User:Haochun zhou/sandbox
Computational Economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems. Some of these areas are unique, while others established areas of economics by allowing robust data analytics and solutions of problems that would be arduous to research without computers and associated numerical methods.[2]
Computational methods have been applied in various fields of economics research, including but not limiting to:
Econometrics: Non-parametric approaches, Semi-parametric approaches, and Machine Learning.
Dynamic Systems Modeling: Optimization, Dynamic stochastic general equilibrium modeling, and Agent-based modeling[3]
History[edit]
Computational economics developed concurrently with the mathematization of the field. During the early 20 century, pioneers such as Jan Tinbergen and Ragnar Frisch advanced the computerization of economics and the growth of econometrics. As a result of advancements in Econometrics, regression models, hypothesis testing, and other computational statistical methods became widely adopted in economic research. On the theoretical front, complex macroeconomic models, including the Real Business Cycle (RBC) model and Dynamic Stochastic General Equilibrium (DSGE) models have propelled the development and application of numerical solution methods that rely heavily on computation. In the 21th century, the development of computational algorithms created new means for computational methods to interact with economic research. Innovative approaches such as machine learning models and agent-based modeling have been actively explored in different areas of economic research, offering economists an expanded toolkit that frequently differs in character from traditional methods.
Applications[edit]
Agent based modelling[edit]
main article: Agent based model
Computational economics uses computer-based economic modeling to solve analytically and statistically formulated economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[4] As such, it is an economic adaptation of the complex adaptive systems paradigm.[5] Here the "agent" refers to "computational objects modeled as interacting according to rules," not real people.[6] Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimization by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[7] including game-theoretical contexts.[8] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is to test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time.[9]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast, complex, unstructured data sets. Various machine learning methods such as the kernel method and random forest have been developed and utilized in data-mining and statistical analysis. These models provide superior classification, predictive capabilities, flexibility compared to traditional statistical models, such as that of the STAR method. Other methods, such as causal machine learning and causal tree, provide distinct advantages, including inference testing.
There are notable advantages and disadvantages of utilizing machine learning tools in economic research. In economics, a model is selected and analyzed at once. The economic research would select a model based on principle, then test/analyze the model with data, followed by cross-validation with other models. On the other hand, machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validates, estimates, and compares various models concurrently. This process may yield more robust estimates than those of the traditional ones.
Traditional economics partially normalize the data based on existing principles, while machine learning presents a more positive/empirical approach to model fitting. Although Machine Learning excels at classification, predication and evaluating goodness of fit, many models lack the capacity for statistical inference, which are of greater interest to economic researchers. Machine learning models' limitations means that economists utilizing machine learning would need to develop strategies for robust, statistical causal inference, a core focus of modern empirical research. For example, economics researchers might hope to identify confounders, confidence intervals, and other parameters that are not well-specified in Machine Learning algorithms.[10]
Machine learning may effectively enable the development of more complicated heterogeneous economic models. Traditionally, heterogeneous models required extensive computational work. Since heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints, optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent).[11] The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[12]
The adoption and implementation of neural networks, deep learning in the field of computational economics may reduce the redundant work of data cleaning and data analytics, significantly lowering the time and cost of large scale data analytics and enabling researchers to collect, analyze data on a great scale.[13] This would encourage economic researchers to explore new modeling methods. In addition, reduced emphasis on data analysis would enable researchers to focus more on subject matters such as causal inference, confounding variables, and realism of the model. Under the proper guidance, machine learning models may accelerate the process of developing accurate, applicable economics through large scale empirical data analysis and computation.[14]
Dynamic Stochastic General Equilibrium (DSGE) model[edit]
main article: DSGE model
Dynamic modeling methods are frequently adopted in macroeconomic research to simulate economic fluctuations and test for the effects of policy changes. The DSGE one class of dynamic models relying heavily on computational techniques and solutions. DSGE models utilize micro-founded economic principles to capture characteristics of the real world economy in an environment with intertemporal uncertainty. Given their inherent complexity, DSGE models are in general analytically intractable, and are usually implemented numerically using computer software. One major advantage of DSGE models is that they facilitate the estimation of agents’ dynamic choices with flexibility. However, many scholars have criticized DSGE models for their reliance on reduced-form assumptions that are largely unrealistic.
Computational tools and programming languages[edit]
Utilizing computational tools in economic research has been the norm and foundation for a long time. Computational tools for economics include a variety of computer software that facilitate the execution of various matrix operations (e.g. matrix inversion) and the solution of systems of linear and nonlinear equations. Various programming languages are utilized in economic research for the purpose of data analytics and modeling. Following is a typical listing of programming languages used in computational economics research:
C++, MATLAB, Julia (programming language), Python (programming language), R (programming language), Stata
Among these programming languages, C++ as a compiled language performs the fastest, while Python as an interpreted language is the slowest. MATLAB, Julia, and R achieve a balance between performance and interpretability. As an early statistical analytics software, Stata was the most conventional programming language option. Economists embraced Stata as one of the most popular statistical analytics programs due to its breadth, accuracy, flexibility, and repeatability.
Journals[edit]
The following journals specialise in computational economics: ACM Transactions on Economics and Computation,[15] Computational Economics,[16] Journal of Applied Econometrics,[17] Journal of Economic Dynamics and Control[18] and the Journal of Economic Interaction and Coordination.[19]
End[edit]
_____
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems, Some of these areas are unique, while others extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious to study without computers and associated numerical methods.[20]
Computational methods in Econometrics: such as non-parametric approaches, semi-parametric approaches, markov processes.
Agent-based method[21]: such as machine learning, evolutionary algorithms, neural network modeling.
Computational methods of dynamic systems: such as optimization, general-equilibrium,[22] equilibrium modeling, Dynamic stochastic general equilibrium.
Computational tools for the design of automated internet markets, and programming tool specifically designed for computational economics and the teaching of computational economics.
History[edit]
Computational economics was development simultaniously with the dvelopment of econometrics. Jan Tinbergen, and Ragnar Frisch transformed economics from a verbal to a mathematical discipline. He developed the first macroeconomic model in the 1930s, mathematically connecting data from the whole economy. He conducted a quantitative research of the US economy's macroeconomic linkages and produced a two-volume book titled Statistical Testing of Business Cycles in 1939. With the development of Econometrics, regression models, hypothesis testing and other form of statistical methods became widely adopted in economic researchs. Researchers combine the economics principles and data analysis to formulat, test and cross examine models. In the 21th century, the development of computer and computational algorithm created new means which computational methods may interact with economic research: such as the development of deep neural network and computational language, enabling economist to process and analyze large sets of data efficiently.
Applications[edit]
Agent based computational economics[edit]
Computational economics uses computer-based economic modelling for the solution of analytically and statistically- formulated economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[23] As such, it is an economic adaptation of the complex adaptive systems paradigm.[24] Here the "agent" refers to "computational objects modeled as interacting according to rules," not real people.[21] Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimisation by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[25] including game-theoretical contexts.[26] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is to test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time.[27]
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model are commonly used in empirical economics research for their flexibility in modeling both homogeneous and heterogeneous functions. Although flexible, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.[28]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast, complex, confounding data sets. This quality, combined with its excellent computational capabilities made it an ideal method to model heterogeneity in economics under ACE, and in solving optimality functions. Various machine learning methods such as the kernel methods, random forests methods were utilized in heterogeneous analysis, these models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enabled researchers to test for inference.
The adoption and implementation of neural networks, deep learning in the field of computational economics[29] may finally combine robust data analysis, impactful empirical analysis and critical inference. Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.”[30]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In economic, a model is selected and analyzed at once. The economic research would select a model based on principle, then test/analyze the model with data, followed by cross-validation with other models. Machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validate, estimates and compares various models, this process may yield more robust estimates. Traditional economics partially normalize the data based on existing principles, while machine learning presents a more positive approach to model fitting. Although Machine learning models excel at classification, predication and evaluating goodness of fit, it lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for its estimated effects. Its limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.[31]
Machine learning may effectively enable the development of more complicated heterogeneou economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”[32], optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[33]
Dynamic Stochastic General Equilibrium model[edit]
Computational economics is used in dynamic modeling to test for the effects of policy changes, specifcally monetary policy changes' effect on the economy.[34] The terms of DSGE, (Dynamic, Stochastic, Equilibrium) attempt to utilize macro economics principles to capture characterisitics of real work economy. However, many scholars have critisized the DSGE model for its overly reliant on general economics principles. Scholars argues that by making the preassumption of an equilibrium, the model fails to caputre the dynamic and stoachstic aspects of the economy. volution of agents choices, stocks, financial assets, economic. Computational economics facilitates DSGE model in estimating the dynmaic choices of agents especially under heterogeneous setting.
Computational tools and programming languages[edit]
Computational solution tools include for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing languages are utilized in economic research for the purpose of data analystics and modeling.
C++, MATLAB, Julia (programming language), Python (programming language), R (programming language), Stata
Among the various programming languages, C++ as a compiled language performs the fastest. Python as an interpreted language is the slowest. While MATLAB, Julia and R balances between performance and interpretability.
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns, Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the base and habit of utilizing computational software in economic research.
Journals[edit]
The following journals specialise in computational economics: ACM Transactions on Economics and Computation,[35] Computational Economics,[1] Journal of Applied Econometrics,[36] Journal of Economic Dynamics and Control[37] and the Journal of Economic Interaction and Coordination.[38]
[one overall comment: the spelling of "Markov" should probably be capitalized throughout]
PEER REIVEW BY DON ASSAMONGKOL (see instructions here)[edit]
(1) Summary of wikipedia entry
(2) Untouched version below
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems: Computational methods in econometrics such as non-parametric approaches, semi-parametric approaches, markov processes. Agent-based method,[21] machine learning, evolutionary algorithms, neural network modeling. Computational methods of dynamic systems such as optimization, general-equilibrium,[22] equilibrium modeling, Dynamic stochastic general equilibrium. Computational tools for the design of automated internet markets, programming tool specifically designed for computational economics and the teaching of computational economics.
Some of these areas are unique, while others extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious to study without computers and associated numerical methods.[20]
Agent based computational economics[edit]
Computational economics uses computer-based economic modelling for the solution of analytically and statistically- formulated economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[23] As such, it is an economic adaptation of the complex adaptive systems paradigm.[24] Here the "agent" refers to "computational objects modeled as interacting according to rules," not real people.[21] Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimisation by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[25] including game-theoretical contexts.[26] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is "to ... test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time, with each research building on the work before."[27]
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model are commonly used in empirical economics research for their flexibility in modeling both homogeneous and heterogeneous functions. Although flexible, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.[28]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast, complex, confounding data sets. This quality, combined with its excellent computational capabilities made it an ideal method to model heterogeneity in economics under ACE, and in solving optimality functions. Various machine learning methods such as the kernel methods, random forests methods were utilized in heterogeneous analysis, these models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enabled researchers to test for inference.
The adoption and implementation of neural networks, deep learning in the field of computational economics[29] may finally combine robust data analysis, impactful empirical analysis and critical inference. Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.”[30]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In economic, a model is selected and analyzed at once. The economic research would select a model based on principle, then test/analyze the model with data, followed by cross-validation with other models. Machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validate, estimates and compares various models, this process may yield more robust estimates. Traditional economics partially normalize the data based on existing principles, while machine learning presents a more positive approach to model fitting. Although Machine learning models excel at classification, predication and evaluating goodness of fit, it lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for its estimated effects. Its limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.[31]
Machine learning may effectively enable the development of more complicated heterogeneous economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”[32], optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[33]
Computational tools and programming languages[edit]
Computational solution tools include for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing languages are utilized in economic research for the purpose of data analystics and modeling.
Among the various programming languages, C++ as a compiled language performs the fastest. Python as an interpreted language is the slowest. While MATLAB, Julia and R balances between performance and interpretability.
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns, Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the base and habit of utilizing computational software in economic research.
Journals[edit]
The following journals specialise in computational economics: ACM Transactions on Economics and Computation,[35] Computational Economics,[1] Journal of Applied Econometrics,[36] Journal of Economic Dynamics and Control[37] and the Journal of Economic Interaction and Coordination.[38]
(3) Revised version with my edits
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems: Computational methods in econometrics such as non-parametric approaches, semi-parametric approaches, markov processes. Agent-based method,[21] machine learning, evolutionary algorithms, neural network modeling. Computational methods of dynamic systems such as optimization, general-equilibrium,[22] equilibrium modeling, Dynamic stochastic general equilibrium. Computational tools for the design of automated internet markets, programming tool specifically designed for computational economics and the teaching of computational economics.
[Maybe it would be more clear to list out the above as bullet points? Eg. - Computational methods: semi-parametric approach, markov, - Agent-based method: ML, evolutionary algos, etc. ]
[Right now the above paragraph is grammatically incorrect I think since you don't have a verb for every sentence]
Some of these areas are unique, while others extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious to study without computers and associated numerical methods.[20]
Agent based computational economics[edit]
Computational economics uses computer-based economic modelling for the solution of analytically and statistically- formulated economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[23] As such, it is an economic adaptation of the complex adaptive systems paradigm.[24] Here the "agent" refers to "computational objects modeled as interacting according to rules," not real people.[21] Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimisation by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[25] including game-theoretical contexts.[26] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is "to ... test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time, with each research building on the work before."[27]
[Nice explanation. I would personally move the explanation of agents higher up in the paragraph before you introduce the ACE paradigm]
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model are commonly used in empirical economics [What is empirical economics? Maybe backlink this] research for their flexibility in modeling both homogeneous and heterogeneous functions. Although flexible, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.[28]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast, complex, confounding data sets. This quality, combined with its excellent computational capabilities made it an ideal method to model heterogeneity in economics under ACE, and in solving optimality functions. Various machine learning methods such as the kernel methods, random forests methods were utilized in heterogeneous analysis, these models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enabled researchers to test for inference.
The adoption and implementation of neural networks, deep learning in the field of computational economics[29] may finally combine robust data analysis, impactful empirical analysis and critical inference. Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.”[30] [Who said this? Would be good to explicitly name author so words can bear more weight ]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In economic[s], a model is selected and analyzed at once. The economic research[er] would select a model based on principle, then test/analyze the model with data, followed by cross-validation with other models. Machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validate[s], estimates and compares various models, this process may yield more robust estimates. Traditional economics partially normalize the data based on existing principles, while machine learning presents a more positive approach to model fitting. [positive here is a technical term, maybe explain it] Although Machine learning models excel at classification, predication and evaluating goodness of fit, it lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for its estimated effects. Its limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.[31]
Machine learning may effectively enable the development of more complicated heterogeneous economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”[32], optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[33]
Computational tools and programming languages[edit]
Computational solution tools include for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing languages are utilized in economic research for the purpose of data analystics and modeling.
Among the various programming languages, C++ as a compiled language performs the fastest. Python as an interpreted language is the slowest. While MATLAB, Julia and R balances between performance and interpretability.
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns, Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the base and habit of utilizing computational software in economic research.
Journals[edit]
The following journals specialise in computational economics: ACM Transactions on Economics and Computation,[35] Computational Economics,[1] Journal of Applied Econometrics,[36] Journal of Economic Dynamics and Control[37] and the Journal of Economic Interaction and Coordination.[38]
Peer Review by Jack Casey[edit]
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems: Computational methods in econometrics such as non-parametric approaches, semi-parametric approaches, markov processes. Agent-based method,[21] machine learning, evolutionary algorithms [Hey, that's my topic!], neural network modeling. Computational methods of dynamic systems such as optimization, general-equilibrium,[22] equilibrium modeling, Dynamic stochastic general equilibrium. Computational tools for the design of automated internet markets, programming tool specifically designed for computational economics and the teaching of computational economics. [You are listing topics, but the periods makes me think of them as incomplete sentences. I would use semicolons if you want a structured list, and make sure that you include phrases like "like", "such as". So it would read something like "this subject encompasses ...: ______ such as A, B, and C; _____s like A, B, and C; and _____s like A and B". Even so, it may be better to forgo a structured list for several full sentences.]
Some of these areas are unique, while others extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious to study without computers and associated numerical methods.[20]
Agent based computational economics[edit]
Computational economics uses computer-based economic modelling for the solution of analytically and statistically- formulated [get ride of that space after the hyphen] economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[23] As such, it is an economic adaptation of the complex adaptive systems paradigm.[24] Here the "agent" refers to "computational objects modeled as interacting according to rules," [I don't like quotes here. I think you almost always want to paraphrase instead] not real people.[21] Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimisation by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[25] including game-theoretical contexts.[26] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is "to ... test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time, with each research building on the work before." [Again, quotes scare me][27]
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model are commonly used in empirical economics research for their flexibility in modeling both homogeneous and heterogeneous functions. Although flexible, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.[28]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve [<-- word choice? handle? utilize? process?] vast, complex, confounding data sets. This quality, combined with its excellent computational capabilities make it an ideal method to model heterogeneity in economics under ACE, and in solving optimality functions. Various machine learning methods such as the kernel methods and random forests methods were utilized in heterogeneous analysis, [and? as?] these models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enable researchers to test for inference.
The adoption and implementation of neural networks, deep learning in the field of computational economics[29] [I would move the citation to end of sentence] may finally [weird tone; is this neutral?] combine robust data analysis, impactful empirical analysis and critical inference. Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.” [Quotes scare me][30]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In [computational?] economics, a model is selected and analyzed at once. The economic research would select a model based on principle, then test or analyze the model with data, followed by cross-validation with other models. Machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validate [-s], estimates and compares various models, this process may yield more robust estimates. [<-- I might restructure this sentence] Traditional economics partially normalizes the data based on existing principles, while machine learning presents a more positive approach to model fitting. Although Machine learning models excel at classification, predication and evaluating goodness of fit, they lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for their estimated effects. Its limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.[31]
Machine learning may effectively enable the development of more complicated heterogeneous economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”[32] [Quotes scare me], optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[33]
Computational tools and programming languages[edit]
Computational solution tools include for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing languages are utilized in economic research for the purpose of data analytics and modeling.
Among the various programming languages, C++ as a compiled language performs the fastest. Python as an interpreted language is the slowest. While MATLAB, Julia and R balances between performance and interpretability. [I agree, but source? Also, is comparing the languages here relevant?]
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns, Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the base and habit of utilizing computational software in economic research.[This section kind of comes out of nowhere? It seems off topic.]
Journals[edit]
The following journals specialise in computational economics: ACM Transactions on Economics and Computation,[35] Computational Economics,[1] Journal of Applied Econometrics,[36] Journal of Economic Dynamics and Control[37] and the Journal of Economic Interaction and Coordination.[38] [I think this section is actually quite useful. Good idea!]
My Thoughts[edit]
Overall, this is a good article. I feel well informed on the subject, and it provides a very good summary of the field with lots of specific examples of sub areas and technologies.
I caught a some grammar mistakes that you should fix. You do use a lot of quotes, which (as I have noted) scare me. I think quotes are discouraged on Wikipedia except in specific circumstances, so I would rather paraphrase.
I would like to acknowledge that my comments are my opinion, and I can certainly be wrong, so be sure to take my advice with a grain of salt.
Strong Work
-- JackCasey067 (talk) 19:43, 7 May 2022 (UTC)
Edits by Jackie Caraveo[edit]
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems: Computational methods in econometrics such as non-parametric approaches, semi-parametric approaches, markov processes. Agent-based method,[21] machine learning, evolutionary algorithms, neural network modeling. Computational methods of dynamic systems such as optimization, general-equilibrium,[22] equilibrium modeling, Dynamic stochastic general equilibrium. Computational tools for the design of automated internet markets, programming tool specifically designed for computational economics and the teaching of computational economics.
Some of these areas are unique, while others extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious to study without computers and associated numerical methods.[20]
[This section seems a bit disorganized because of the awkward sentences. One way to improve this section is to combine sentences about the different approaches and models.]
Agent based computational economics[edit]
Computational economics uses computer-based economic modelling for the solution of analytically and statistically- formulated economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[23] As such, it is an economic adaptation of the complex adaptive systems paradigm.[24] Here the "agent" refers to "computational objects modeled as interacting according to rules," not real people.[21] Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimisation by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[25] including game-theoretical contexts.[26] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is "to ... test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time, with each research building on the work before."[27]
[In this section, make sure to paraphrase the sections where there are direct quotes.]
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model are commonly used in empirical economics research for their flexibility in modeling both homogeneous and heterogeneous functions. Although flexible, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.[28] [I would be careful on adding statements that point to biases since it can be classified as an opinion, not something factual.]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast, complex, confounding data sets. [This sentence is a bit wordy. Reword to: Machine learning models present a method to resolve complex data sets.] This quality, combined with its excellent computational capabilities make it an ideal method to model heterogeneity in economics under ACE, and in solving optimality functions. Various machine learning methods such as the kernel methods,[and] random forests methods were utilized in heterogeneous analysis. These models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enabled researchers to test for inference.
The adoption and implementation of neural networks, [and] deep learning in the field of computational economics[29] [Move this citation to the end of the sentence] may finally combine robust data analysis, impactful empirical analysis and critical inference.
Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.”[30] [This sentence is worded in a way where it's conveying a bias, also there's a direct quote instead of paraphrasing. I suggest rewording the sentence so it's not as biased. The sentence can also be removed since I don't think it's too important in this section. Also I could be wrong, please take it with a grain of salt.]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In economic [economic?], a model is selected and analyzed at once. The economic research would select a model based on principle, then test/analyze the model with data, followed by cross-validation with other models. Machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validate, estimates and compares various models, this process may yield more robust estimates. Traditional economics partially normalize the data based on existing principles, while machine learning presents a more positive approach to model fitting. Although Machine learning models excel at classification, predication and evaluating goodness of fit, it lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for its estimated effects. Its limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.[31] [Make sure to paraphrase, not quote]
Machine learning may effectively enable the development of more complicated heterogeneous economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”[32], optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[33] [Same here, make sure to paraphrase, not quote]
Computational tools and programming languages[edit]
Computational solution tools include for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing languages are utilized in economic research for the purpose of data analystics and modeling.
Among the various programming languages, C++ as a compiled language performs the fastest. Python as an interpreted language is the slowest. While MATLAB, Julia and R balances between performance and interpretability.
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns, Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the base and habit of utilizing computational software in economic research.
[I think this section seems a little out of place. What's the goal of including information about the different programming languages?]
Notes: Hi! I think this a great start! The main thing that I would recommend is to paraphrase instead of using direct quotes. There are also some sections that seem unnecessary, but I could be wrong. I look forward to seeing the final product!
Journals[edit]
The following journals specialise in computational economics: ACM Transactions on Economics and Computation,[35] Computational Economics,[1] Journal of Applied Econometrics,[36] Journal of Economic Dynamics and Control[37] and the Journal of Economic Interaction and Coordination.[38]
Peer Review by Chad Berkich[edit]
All potential edits will be in italics and bolded to make them clearly standout. Comments are bolded and in brackets [example].
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science. This subject encompasses computational modeling of economic systems: Computational methods in econometrics such as non-parametric approaches, semi-parametric approaches, Markov processes. Agent-based method, machine learning, evolutionary algorithms, and neural network modeling. Computational methods of dynamic systems such as optimization, general-equilibrium, equilibrium modeling, and Dynamic stochastic general equilibrium. Computational tools for the design of automated internet markets, programming tool specifically designed for computational economics and the teaching of computational economics. [Please fix the grammar in the proceeding sentence. I am really struggling with what you are trying to say because of the grammar to the extent that I can't actually suggest edits to fix it. I apologize for that, but please make sure that the grammar makes sense.]
Some of these areas are unique to the field, while others extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious without computers and by using associated numerical methods. [My "by using" edit could be wrong as far as meaning. If it is, please correct the grammar to reflect that, as it is currently makes me think that another verb needs to be added there.]
Agent-[adding a dash here because you later say it has a dash]based computational economics[edit]
Computational economics uses computer-based economic modelling for the solution of analytically- and statistically-formulated economic problems [As far as I am aware, the dashes I have added is the convention. However, I could be wrong.]. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[There is a lot of information in this sentence to the extent that I am having trouble understanding it. I would suggest breaking it up into two or three. Also, I am confused by the introduction of a research program at the start. Clarification here would help.] As such, it [What is it here?] is an economic adaptation of the complex adaptive systems paradigm. The "agent" refers to "computational objects modeled as interacting according to rules," not real people. Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimization [note the z has been changed] by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces, including game-theoretical contexts. Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is "to ... test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time, with each research building on the work before."
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model, [note the comma before here] are commonly used in empirical economics research for their flexibility in modeling both homogeneous and heterogeneous functions. However, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast and complex [the confounding part was removed to remove subjectivity] data sets. This quality, combined with its ["excellent" removed for subjectivity] computational capabilities, [note the comma before here] made it an ideal method to model heterogeneity in economics under ACE and to solve optimality functions. Various machine learning methods such as the kernel method and random forests method [note the "s" were removed from "method"] were utilized in heterogeneous analysis. [note the period before here] These models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enabled researchers to test for inference.
The adoption and implementation of neural networks and deep learning in the field of computational economics may ['finally' removed for subjectivity] combine robust data analysis, impactful empirical analysis and critical inference. Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.” [Quote needs more context and probably credit to who said it.]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In economics, a model is selected and analyzed at once [What does at once mean here? I don't have a background in econ, so this may be a misunderstanding on my end, but its kind of confusing]. The economic researcher [Is this edit accurate? If not, it needs to be fixed.] selects a model based on principle, then tests/analyzes the model with data, followed by cross-validation with other models. In contrast, machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validates [note that an s was added to the end of 'validates'], estimates and compares various models. [note the period before here] This process may yield more robust estimates. Traditional economics partially normalizes the data based on existing principles, while machine learning presents a more positive [what does positive mean here? Context would probably be helpful] approach to model fitting. Although machine learning models excel at classification, predication and evaluating accuracy of fit, it lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for their estimated effects. These limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.
Machine learning may effectively enable the development of more complicated heterogeneous economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”, optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.
Computational tools and programming languages[edit]
Computational solution tools [I would suggest explaining what these tools are, and then going into examples of what they do] include for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing languages are utilized in economic research for the purpose of data analytics and modeling. These include the following:
Among these various programming languages, C++ performs the fastest as it is a compiled language [moved the clause for flow]. Python, in contrast, performs the slowest as it is an interpreted language [moved the clause for flow]. [Remove "While" here] MATLAB, Julia and R fall in between C++ and Python on performance speed and interpretability [this may not be the correct word to use here].
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. ["Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns," I would remove this clause as it doesn't add to the point of the article] Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the basis of utilizing computational software in economic research.
Journals[edit]
The following journals specialize [note the change from 's' to 'z'] in computational economics: ACM Transactions on Economics and Computation, Computational Economics, Journal of Applied Econometrics, Journal of Economic Dynamics and Control and the Journal of Economic Interaction and Coordination.
Overall thoughts[edit]
This is a good article that is very detailed. However, there are times when the detail can bog it down and make it harder to read, so I've suggested where to fix that. Additionally, there are frequent grammar mistakes in the article, that often interfere with clarity. I have marked all that I found but I would recommend proofreading multiple times to make sure any other mistakes are caught.
Peer Review by George Afentakis[edit]
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems: Computational methods in econometrics such as non-parametric approaches, semi-parametric approaches, markov processes. Agent-based method,[21] machine learning, evolutionary algorithms, neural network modeling. Computational methods of dynamic systems such as optimization, general-equilibrium,[22] equilibrium modeling, Dynamic stochastic general equilibrium. Computational tools for the design of automated internet markets, programming tool specifically designed for computational economics and the teaching of computational economics.
[Maybe you should rephrase this section, it is a bit non-clear]
Some of these areas are unique, while others extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious to study without computers and associated numerical methods.[20]
Agent based computational economics[edit]
Computational economics uses computer-based economic modelling for the solution of analytically and statistically- formulated economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[23] As such, it is an economic adaptation of the complex adaptive systems paradigm.[24] Here the "agent" refers to "computational objects modeled as interacting according to rules," not real people.[21] Agents can represent social, biological, and/or physical entities. [The sentences up to this point are very dense with a lot of information that can be hard to understand maybe it would help if you break them up into more sentences.] The theoretical assumption of mathematical optimisation by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[25] including game-theoretical contexts.[26] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is "to ... test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time, with each research building on the work before."[27]
[Since you are including a direct quote in the last sentence I think it should exactly match the reference i.e. "to test their theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time, with each researcher's work building appropriately on the work that has gone before." If you don't want to use the direct quote maybe paraphrase it a bit and remove most of the quotes.]
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model are commonly used in empirical economics research for their flexibility in modeling both homogeneous and heterogeneous functions. Although flexible, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.[28]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast, complex, confounding data sets. This quality, combined with its excellent computational capabilities made it an ideal method to model heterogeneity in economics under ACE, and in solving optimality functions. Various machine learning methods such as the kernel methods, random forests methods were utilized in heterogeneous analysis, these models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enabled researchers to test for inference.
The adoption and implementation of neural networks, deep learning in the field of computational economics[29] may finally combine robust data analysis, impactful empirical analysis and critical inference. Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.”[30]
[Don't have a sentence that has only the quote. Either connect the quote with the sentence before or paraphrase it. Maybe you can write ... allowing economic researchers (or just economists) to focus on empirical work since they need to only define the original model and the parameter of interest]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In economic[s], a model is selected and analyzed at once. [I'm a little confused what at once means and this sentence seems disconnected from the next one maybe adding some transitioning words could help and make it more clear.] The economic research would select a model based on principle, then test/analyze the model with data, followed by cross-validation with other models. Machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validate, estimates and compares various models, this process may yield more robust estimates. Traditional economics partially normalize the data based on existing principles, while machine learning presents a more positive approach to model fitting. Although Machine learning models excel at classification, predication and evaluating goodness of fit, it lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for its estimated effects. Its limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.[31] [Again you have a lot of information clear and maybe you should add more context to make it easier to read for a third-party]
Machine learning may effectively enable the development of more complicated heterogeneous economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”[32], optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[33]
Computational tools and programming languages[edit]
Computational solution tools include [maybe add a little more information on what exactly are before writing up examples] for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing languages are utilized in economic research for the purpose of data analystics and modeling.
[I think it looks better to have all the examples on the same line.]
Among the various programming languages, C++ as a compiled language performs the fastest. Python as an interpreted language is the slowest. While MATLAB, Julia and R balances between performance and interpretability.
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns, Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the base and habit of utilizing computational software in economic research.
Journals[edit]
The following journals specialise in computational economics: ACM Transactions on Economics and Computation,[35] Computational Economics,[1] Journal of Applied Econometrics,[36] Journal of Economic Dynamics and Control[37] and the Journal of Economic Interaction and Coordination.[38]
[Overall thoughts: You have done a good work and you have gathered a lot of information. This makes your sentences very dense and maybe hard to read for someone that does not have an econ background. If you rephrase some sentences and make them more clear you will have a very good article. ]
Peer Review by Jack Kim[edit]
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems: Computational methods in econometrics such as non-parametric approaches, semi-parametric approaches, [and] markov processes. Agent-based method,[21] machine learning, evolutionary algorithms, neural network modeling. Computational methods of dynamic systems such as optimization, general-equilibrium,[22] equilibrium modeling, Dynamic stochastic general equilibrium. Computational tools for the design of automated internet markets, programming tool specifically designed for computational economics and the teaching of computational economics [the last two sentences are incomplete sentences].
Some of these areas are unique, while others [i think the use of the word areas and others in this context is ambiguous] extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious to study without computers and associated numerical methods.[20]
Agent based computational economics[edit]
Computational economics uses computer-based economic modelling for [creating] the solution of[for instead of of] analytically and statistically- formulated economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[23] [the previous sentence is unclear and hard to read/ not a complete sentence] As such, it is an economic adaptation of the complex adaptive systems paradigm.[24] Here the "agent" refers to "computational objects modeled as interacting according to rules," not real people.[21] Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimisation by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[25] including game-theoretical contexts.[26] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is "to ... test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time, with each research building on the work before."[27]
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model are commonly used in empirical economics research for their flexibility in modeling both homogeneous and heterogeneous functions. Although flexible, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.[28]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast, complex, confounding data sets. This quality, combined with its excellent computational capabilities made it an ideal method to model heterogeneity in economics under ACE, and in solving optimality functions. Various machine learning methods such as the kernel methods, random forests methods were utilized in heterogeneous analysis, these models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enabled researchers to test for inference. [this sentence is needs to be broken up into two sentences]
The adoption and implementation of neural networks, deep learning in the field of computational economics[29] may finally combine robust data analysis, impactful empirical analysis and critical inference. Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.”[30]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In economic, a model is selected and analyzed at once. The economic research would select a model based on principle, then test/analyze the model with data, followed by cross-validation with other models. Machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validate, estimates and compares various models, this process may yield more robust estimates. Traditional economics partially normalize the data based on existing principles, while machine learning presents a more positive approach to model fitting. Although Machine learning models excel at classification, predication and evaluating goodness of fit, it lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for its estimated effects. Its limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.[31]
Machine learning may effectively enable the development of more complicated heterogeneous economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”[32], optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[33]
Computational tools and programming languages[edit]
Computational solution tools include for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing languages are utilized in economic research for the purpose of data analystics and modeling.
Among the various programming languages, C++ as a compiled language performs the fastest. Python as an interpreted language is the slowest. While MATLAB, Julia and R balances between performance and interpretability.
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns, Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the base and habit of utilizing computational software in economic research.
Journals[edit]
The following journals specialise in computational economics: ACM Transactions on Economics and Computation,[35] Computational Economics,[1] Journal of Applied Econometrics,[36] Journal of Economic Dynamics and Control[37] and the Journal of Economic Interaction and Coordination.[38]
I think the content itself is excellent but sentence structures and grammatical errors throughout the page make it difficult to digest.
Peer Review by Yamato Hart[edit]
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems: Computational methods in econometrics such as non-parametric approaches, semi-parametric approaches, markov [capitalization?]processes. Agent-based method,[21] machine learning, evolutionary algorithms, neural network modeling. Computational methods of dynamic systems such as optimization, general-equilibrium,[22] equilibrium modeling, Dynamic stochastic general equilibrium. Computational tools for the design of automated internet markets, programming tool specifically designed for computational economics and the teaching of computational economics .
Some of these areas are unique, while others ' extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious to study without computers and associated numerical methods.[20]
Agent based computational economics[edit]
Computational economics uses computer-based economic modelling for the solution of analytically and statistically- formulated economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[23] As such, it is an economic adaptation of the complex adaptive systems paradigm.[24] Here the "agent" refers to "computational objects modeled as interacting according to rules," not real people.[21] Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimisation by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[25] including game-theoretical contexts.[26] [The past two sentences are a bit hard to understand] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is "to ... test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time, with each research building on the work before."[27]
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model are commonly used in empirical economics research for their flexibility in modeling both homogeneous and heterogeneous functions. Although flexible, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.[28]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast, complex, confounding data sets. This quality, combined with its excellent computational capabilities made it an ideal method to model heterogeneity in economics under ACE, and in solving optimality functions. Various machine learning methods such as the kernel methods, random forests methods [erroneous comma?]were utilized in heterogeneous analysis, these models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enabled researchers to test for inference. '
The adoption and implementation of neural networks, deep learning in the field of computational economics[29] may finally combine robust data analysis, impactful empirical analysis and critical inference. Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.”[30]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In economic, [in economics?]a model is selected and analyzed at once. The economic research would select a model based on principle, then test/analyze the model with data, followed by cross-validation with other models. Machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validate, estimates and compares various models, this process may yield more robust estimates. Traditional economics partially normalize the data based on existing principles, while machine learning presents a more positive approach to model fitting. Although Machine learning models excel at classification, predication and evaluating goodness of fit, it lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for its estimated effects. Its limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.[31]
Machine learning may effectively enable the development of more complicated heterogeneous economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”[32], optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[33]
Computational tools and programming languages[edit]
Computational solution tools include for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing [programming?] languages are utilized in economic research for the purpose of data analystics ([analytics?]) and modeling.
Among the various programming languages, C++ as a compiled language performs the fastest. Python as an interpreted language is the slowest. While MATLAB, Julia and R balances between performance and interpretability.
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns, Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the base and habit of utilizing computational software in economic research.
Journals[edit]
The following journals specialise in computational economics: ACM Transactions on Economics and Computation,[35] Computational Economics,[1] Journal of Applied Econometrics,[36] Journal of Economic Dynamics and Control[37] and the Journal of Economic Interaction and Coordination.[38]
References[edit]
- ^ a b c d e f g h i j k l m n o p q Computational Economics. ""About This Journal" and "Aims and Scope."
- ^ • Hans M. Amman, David A. Kendrick, and John Rust, ed., 1996. Handbook of Computational Economics, v. 1, Elsevier. Description Archived 2011-07-15 at the Wayback Machine & chapter-preview links. Archived 2020-04-06 at the Wayback Machine • Kenneth L. Judd, 1998. Numerical Methods in Economics, MIT Press. Links to description Archived 2012-02-11 at the Wayback Machine and chapter previews.
- ^ Scott E. Page, 2008. "agent-based models," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract.
- ^ • Scott E. Page, 2008. "agent-based models," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract. • Leigh Tesfatsion, 2006. "Agent-Based Computational Economics: A Constructive Approach to Economic Theory," ch. 16, Handbook of Computational Economics, v. 2, [pp. 831-880]. doi:10.1016/S1574-0021(05)02016-2. • Kenneth L. Judd, 2006. "Computationally Intensive Analyses in Economics," Handbook of Computational Economics, v. 2, ch. 17, pp. 881- 893. Pre-pub PDF. • L. Tesfatsion and K. Judd, ed., 2006. Handbook of Computational Economics, v. 2, Agent-Based Computational Economics, Elsevier. Description Archived 2012-03-06 at the Wayback Machine & and chapter-preview links. • Thomas J. Sargent, 1994. Bounded Rationality in Macroeconomics, Oxford. Description and chapter-preview 1st-page links.
- ^ • W. Brian Arthur, 1994. "Inductive Reasoning and Bounded Rationality," American Economic Review, 84(2), pp. 406-411 Archived 2013-05-21 at the Wayback Machine. • Leigh Tesfatsion, 2003. "Agent-based Computational Economics: Modeling Economies as Complex Adaptive Systems," Information Sciences, 149(4), pp. 262-268 Archived April 26, 2012, at the Wayback Machine. • _____, 2002. "Agent-Based Computational Economics: Growing Economies from the Bottom Up," Artificial Life, 8(1), pp.55-82. Abstract and pre-pub PDF Archived 2013-05-14 at the Wayback Machine.
- ^ Scott E. Page, 2008. "agent-based models," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract.
- ^ • W. Brian Arthur, 1994. "Inductive Reasoning and Bounded Rationality," American Economic Review, 84(2), pp. 406-411 Archived 2013-05-21 at the Wayback Machine. • John H. Holland and John H. Miller (1991). "Artificial Adaptive Agents in Economic Theory," American Economic Review, 81(2), pp. 365-370 Archived 2011-01-05 at the Wayback Machine. • Thomas C. Schelling, 1978 [2006]. Micromotives and Macrobehavior, Norton. Description Archived 2017-11-02 at the Wayback Machine, preview. • Thomas J. Sargent, 1994. Bounded Rationality in Macroeconomics, Oxford. Description and chapter-preview 1st-page links.
- ^ • Joseph Y. Halpern, 2008. "computer science and game theory," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract. • Yoav Shoham, 2008. "Computer Science and Game Theory," Communications of the ACM, 51(8), pp. 75-79 Archived 2012-04-26 at the Wayback Machine. • Alvin E. Roth, 2002. "The Economist as Engineer: Game Theory, Experimentation, and Computation as Tools for Design Economics," Econometrica, 70(4), pp. 1341–1378 Archived 2004-04-14 at the Wayback Machine.
- ^ Leigh Tesfatsion, 2006. "Agent-Based Computational Economics: A Constructive Approach to Economic Theory," ch. 16, Handbook of Computational Economics, v. 2, sect. 5, p. 865 [pp. 831-880]. doi:10.1016/S1574-0021(05)02016-2.
- ^ "The Impact of Machine Learning on Economics", The Economics of Artificial Intelligence, University of Chicago Press, pp. 507–552, 2019, retrieved 2022-05-05
- ^ Jesus, Browning, Martin Carro, (2006). Heterogeneity and microeconometrics modelling. CAM, Centre for Applied Microeconometrics. OCLC 1225293761.
{{cite book}}: CS1 maint: extra punctuation (link) CS1 maint: multiple names: authors list (link) - ^ Charpentier, Arthur; Élie, Romuald; Remlinger, Carl (2021-04-23). "Reinforcement Learning in Economics and Finance". Computational Economics. doi:10.1007/s10614-021-10119-4. ISSN 1572-9974.
- ^ Farrell, Max H.; Liang, Tengyuan; Misra, Sanjog (2021). "Deep Neural Networks for Estimation and Inference". Econometrica. 89 (1): 181–213. doi:10.3982/ecta16901. ISSN 0012-9682.
- ^ "Deep learning for individual heterogeneity: an automatic inference framework". 2021-07-27.
{{cite journal}}: Cite journal requires|journal=(help) - ^ "ACM Teac".
- ^ Computational Economics. ""About This Journal" and "Aims and Scope."
- ^ "Journal of Applied Econometrics". Wiley Online Library. 2011. doi:10.1002/(ISSN)1099-1255. Retrieved October 31, 2011.
- ^ Journal of Economic Dynamics and Control, including Aims & scope link. For a much-cited overview and issue, see: • Leigh Tesfatsion, 2001. "Introduction to the Special Issue on Agent-based Computational Economics," Journal of Economic Dynamics & Control, pp. 281-293. • [Special issue], 2001. Journal of Economic Dynamics and Control, Agent-based Computational Economics (ACE). 25(3-4), pp. 281-654. Abstract/outline links[permanent dead link].
- ^ "Journal of Economic Interaction and Coordination". springer.com. 2011. Retrieved October 31, 2011.
- ^ a b c d e f g h • Hans M. Amman, David A. Kendrick, and John Rust, ed., 1996. Handbook of Computational Economics, v. 1, Elsevier. Description Archived 2011-07-15 at the Wayback Machine & chapter-preview links. Archived 2020-04-06 at the Wayback Machine • Kenneth L. Judd, 1998. Numerical Methods in Economics, MIT Press. Links to description Archived 2012-02-11 at the Wayback Machine and chapter previews.
- ^ a b c d e f g h i j k l m n o p Scott E. Page, 2008. "agent-based models," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract.
- ^ a b c d e f g h The New Palgrave Dictionary of Economics, 2008. 2nd Edition: • "computation of general equilibria" by Herbert E. Scarf. Abstract. • "computation of general equilibria (new developments)" by Felix Kubler. Abstract.
- ^ a b c d e f g h • Scott E. Page, 2008. "agent-based models," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract. • Leigh Tesfatsion, 2006. "Agent-Based Computational Economics: A Constructive Approach to Economic Theory," ch. 16, Handbook of Computational Economics, v. 2, [pp. 831-880]. doi:10.1016/S1574-0021(05)02016-2. • Kenneth L. Judd, 2006. "Computationally Intensive Analyses in Economics," Handbook of Computational Economics, v. 2, ch. 17, pp. 881- 893. Pre-pub PDF. • L. Tesfatsion and K. Judd, ed., 2006. Handbook of Computational Economics, v. 2, Agent-Based Computational Economics, Elsevier. Description Archived 2012-03-06 at the Wayback Machine & and chapter-preview links. • Thomas J. Sargent, 1994. Bounded Rationality in Macroeconomics, Oxford. Description and chapter-preview 1st-page links.
- ^ a b c d e f g h • W. Brian Arthur, 1994. "Inductive Reasoning and Bounded Rationality," American Economic Review, 84(2), pp. 406-411 Archived 2013-05-21 at the Wayback Machine. • Leigh Tesfatsion, 2003. "Agent-based Computational Economics: Modeling Economies as Complex Adaptive Systems," Information Sciences, 149(4), pp. 262-268 Archived April 26, 2012, at the Wayback Machine. • _____, 2002. "Agent-Based Computational Economics: Growing Economies from the Bottom Up," Artificial Life, 8(1), pp.55-82. Abstract and pre-pub PDF Archived 2013-05-14 at the Wayback Machine.
- ^ a b c d e f g h • W. Brian Arthur, 1994. "Inductive Reasoning and Bounded Rationality," American Economic Review, 84(2), pp. 406-411 Archived 2013-05-21 at the Wayback Machine. • John H. Holland and John H. Miller (1991). "Artificial Adaptive Agents in Economic Theory," American Economic Review, 81(2), pp. 365-370 Archived 2011-01-05 at the Wayback Machine. • Thomas C. Schelling, 1978 [2006]. Micromotives and Macrobehavior, Norton. Description Archived 2017-11-02 at the Wayback Machine, preview. • Thomas J. Sargent, 1994. Bounded Rationality in Macroeconomics, Oxford. Description and chapter-preview 1st-page links.
- ^ a b c d e f g h • Joseph Y. Halpern, 2008. "computer science and game theory," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract. • Yoav Shoham, 2008. "Computer Science and Game Theory," Communications of the ACM, 51(8), pp. 75-79 Archived 2012-04-26 at the Wayback Machine. • Alvin E. Roth, 2002. "The Economist as Engineer: Game Theory, Experimentation, and Computation as Tools for Design Economics," Econometrica, 70(4), pp. 1341–1378 Archived 2004-04-14 at the Wayback Machine.
- ^ a b c d e f g h Leigh Tesfatsion, 2006. "Agent-Based Computational Economics: A Constructive Approach to Economic Theory," ch. 16, Handbook of Computational Economics, v. 2, sect. 5, p. 865 [pp. 831-880]. doi:10.1016/S1574-0021(05)02016-2.
- ^ a b c d e f g h Demetrescu, Matei; Leppin, Julian S.; Reitz, Stefan (2020-06-11). "Homogeneous vs. heterogeneous transition functions in panel smooth transition regressions". Econometric Reviews. 40 (2): 177–196. doi:10.1080/07474938.2020.1773666. ISSN 0747-4938.
- ^ a b c d e f g h Farrell, Max H.; Liang, Tengyuan; Misra, Sanjog (2021). "Deep Neural Networks for Estimation and Inference". Econometrica. 89 (1): 181–213. doi:10.3982/ecta16901. ISSN 0012-9682.
- ^ a b c d e f g h "Deep learning for individual heterogeneity: an automatic inference framework". 2021-07-27.
{{cite journal}}: Cite journal requires|journal=(help) - ^ a b c d e f g h "The Impact of Machine Learning on Economics", The Economics of Artificial Intelligence, University of Chicago Press, pp. 507–552, 2019, retrieved 2022-05-05
- ^ a b c d e f g h Jesus, Browning, Martin Carro, (2006). Heterogeneity and microeconometrics modelling. CAM, Centre for Applied Microeconometrics. OCLC 1225293761.
{{cite book}}: CS1 maint: extra punctuation (link) CS1 maint: multiple names: authors list (link) - ^ a b c d e f g h Charpentier, Arthur; Élie, Romuald; Remlinger, Carl (2021-04-23). "Reinforcement Learning in Economics and Finance". Computational Economics. doi:10.1007/s10614-021-10119-4. ISSN 1572-9974.
- ^ "Oversight hearings on Public Law 93-410, the geothermal energy research, development and demonstration act of 1974: hearing before the Subcommittee on Energy Research, Development and Demonstration of the Committee on Science and Technology, U.S. House of Representatives, second session, January 20, 1976". 1976.
{{cite journal}}: Cite journal requires|journal=(help) - ^ a b c d e f g h "ACM Teac".
- ^ a b c d e f g h "Journal of Applied Econometrics". Wiley Online Library. 2011. doi:10.1002/(ISSN)1099-1255. Retrieved October 31, 2011.
- ^ a b c d e f g h Journal of Economic Dynamics and Control, including Aims & scope link. For a much-cited overview and issue, see: • Leigh Tesfatsion, 2001. "Introduction to the Special Issue on Agent-based Computational Economics," Journal of Economic Dynamics & Control, pp. 281-293. • [Special issue], 2001. Journal of Economic Dynamics and Control, Agent-based Computational Economics (ACE). 25(3-4), pp. 281-654. Abstract/outline links[permanent dead link].
- ^ a b c d e f g h "Journal of Economic Interaction and Coordination". springer.com. 2011. Retrieved October 31, 2011.
External links[edit]
- Society for Computational Economics
- Journal of Economic Dynamics and Control - publishes articles on computational economics
- Agent-Based Computational Economics - maintained by Leigh Tesfatsion
- The Use of Agent-Based Models in Regional Science - a study on agent-based models to simulate urban agglomeration
- Computational Economics with Python - a series of lectures
- Computational Finance and Economic Agents
- Journal of Economic Interaction and Coordination - official journal of the Association of Economic Science with Heterogeneous Interacting Agents
- Chair of Economic Policy, University of Bamberg (Germany)
- Repository of public-domain computational solutions
Peer Review by Je Yeong Soh[edit]
Computational economics is an interdisciplinary research discipline that involves computer science, economics, and management science.[1] This subject encompasses computational modeling of economic systems, Some of these areas are unique, while others extend traditional areas of economics by enabling robust data analytics, solving problems that are tedious to study without computers and associated numerical methods.[2]
Computational methods in Econometrics: such as non-parametric approaches, semi-parametric approaches, markov processes.
Agent-based method[3]: such as machine learning, evolutionary algorithms, neural network modeling.
Computational methods of dynamic systems: such as optimization, general-equilibrium,[4] equilibrium modeling, Dynamic stochastic general equilibrium.
Computational tools for the design of automated internet markets, and programming tool specifically designed for computational economics and the teaching of computational economics.
[Could use bullet-based formatting here to make the list more readable.]
History[edit]
Computational economics was development simultaniously with the dvelopment of econometrics. [Typos: simultaneously, development] Jan Tinbergen, and Ragnar Frisch transformed economics from a verbal to a mathematical discipline. He developed the first macroeconomic model in the 1930s, mathematically connecting data from the whole economy. He conducted a quantitative research of the US economy's macroeconomic linkages and produced a two-volume book titled Statistical Testing of Business Cycles in 1939. With the development of Econometrics, regression models, hypothesis testing and other form of statistical methods became widely adopted in economic researchs. Researchers combine the economics principles and data analysis to formulat, test and cross examine models. In the 21th century, the development of computer and computational algorithm created new means which computational methods may interact with economic research: such as the development of deep neural network and computational language, enabling economist to process and analyze large sets of data efficiently. [Several typos.]
Applications[edit]
Agent based computational economics[edit]
Computational economics uses computer-based economic modelling for the solution of analytically and statistically- formulated economic problems. A research program, to that end, is agent-based computational economics (ACE), the computational study of economic processes, including whole economies, as dynamic systems of interacting agents.[5] As such, it is an economic adaptation of the complex adaptive systems paradigm.[6] Here the "agent" refers to "computational objects modeled as interacting according to rules," not real people.[3] Agents can represent social, biological, and/or physical entities. The theoretical assumption of mathematical optimisation by agents in equilibrium is replaced by the less restrictive postulate of agents with bounded rationality adapting to market forces,[7] including game-theoretical contexts.[8] Starting from initial conditions determined by the modeler, an ACE model develops forward through time driven solely by agent interactions. The scientific objective of the method is to test theoretical findings against real-world data in ways that permit empirically supported theories to cumulate over time.[9]
Modeling methods[edit]
Smooth Transition Models, such as Smooth Transition Autoregressive Model (STAR model), Vector Autoregressive Model and Panel Regression Model are commonly used in empirical economics research for their flexibility in modeling both homogeneous and heterogeneous functions. Although flexible, this class of models may yield biased estimates if used under homogeneous assumption, while the true model is heterogeneous.[10]
Machine learning in computational economics[edit]
Machine learning models present a method to resolve vast, complex, confounding data sets. This quality, combined with its excellent computational capabilities made it an ideal method to model heterogeneity in economics under ACE, and in solving optimality functions. Various machine learning methods such as the kernel methods, random forests methods were utilized in heterogeneous analysis, these models have excellent classification and regression capabilities, while methods such as causal machine learning and causal tree enabled researchers to test for inference. [Citations in this paragraph would be helpful.]
The adoption and implementation of neural networks, deep learning in the field of computational economics[11] may finally combine robust data analysis, impactful empirical analysis and critical inference. Deep neural networks may reduce the redundant work of data cleaning and data analytics, allowing economic researchers to focus on empirical work. “The researcher need only define the original model and define the parameter of interest.”[12]
There are notable advantages and disadvantages of utilizing machine learning in economic research. In economic, a model is selected and analyzed at once. The economic research would select a model based on principle, then test/analyze the model with data, followed by cross-validation with other models. Machine learning models have built in “tuning” effects. As the model conducts empirical analysis, it cross-validate, estimates and compares various models, this process may yield more robust estimates. Traditional economics partially normalize the data based on existing principles, while machine learning presents a more positive approach to model fitting. Although Machine learning models excel at classification, predication and evaluating goodness of fit, it lacks the ability to draw inferences, which are of greater interest to economic researchers. Machine learning models also do not provide a clear confidence interval for its estimated effects. Its limitations meant that economists utilizing machine learning would need to develop effective strategies to identify causal inference, confounders, confidence intervals, outcome metrics that reflect the overall objective, and other parameters.[13]
Machine learning may effectively enable the development of more complicated heterogeneou economic models. Traditionally, heterogeneous models required extensive computational work. Since “heterogeneity could be differences in tastes, beliefs, abilities, skills or constraints”[14], optimizing a heterogeneous model is a lot more tedious than the homogeneous approach (representative agent). The development of reinforced learning and deep learning may significantly reduce the complexity of heterogeneous analysis, creating models that better reflect agents’ behaviors in the economy.[15]
Dynamic Stochastic Gneral Equilibrium model[edit]
Computational economics is used in dynamic modeling to test for the effects of policy changes, specifcally monetary policy changes' effect on the economy.[16] The terms of DSGE, (Dynmamic, Stochastic, Equilibrium) attempt to utilize macro economics principles to capture characterisitics of real work economy. However, many scholars have critisized the DSGE model for its overly reliant on general economics principles. Scholars argues that by making the preassumption of an equilibrium, the model fails to caputre the dynamic and stoachstic aspects of the economy. volution of agents choices, stocks, financial assets, economic. Computational economics facilitates DSGE model in estimating the dynmaic choices of agents especially under heterogeneous setting. [Citation would be helpful.]
Computational tools and programming languages[edit]
Computational solution tools include for example software for carrying out various matrix operations (e.g. matrix inversion) and for solving systems of linear and non-linear equations. Various programing languages are utilized in economic research for the purpose of data analystics and modeling.
C++, MATLAB, Julia (programming language), Python (programming language), R (programming language), Stata
Among the various programming languages, C++ as a compiled language performs the fastest. Python as an interpreted language is the slowest. While MATLAB, Julia and R balances between performance and interpretability. [Citation would be helpful.]
As an early statistical analytics software, economists utilized Stata for its breadth, accuracy, extensibility, and repeatability. Whether the aim is to investigate school selection, minimum wage, GDP, or stock market patterns, Stata offers the statistics, graphics, and data management tools necessary to investigate a wide variety of economic issues. It forms the base and habit of utilizing computational software in economic research.
- ^ Computational Economics. ""About This Journal" and "Aims and Scope."
- ^ • Hans M. Amman, David A. Kendrick, and John Rust, ed., 1996. Handbook of Computational Economics, v. 1, Elsevier. Description Archived 2011-07-15 at the Wayback Machine & chapter-preview links. Archived 2020-04-06 at the Wayback Machine • Kenneth L. Judd, 1998. Numerical Methods in Economics, MIT Press. Links to description Archived 2012-02-11 at the Wayback Machine and chapter previews.
- ^ a b Scott E. Page, 2008. "agent-based models," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract.
- ^ The New Palgrave Dictionary of Economics, 2008. 2nd Edition: • "computation of general equilibria" by Herbert E. Scarf. Abstract. • "computation of general equilibria (new developments)" by Felix Kubler. Abstract.
- ^ • Scott E. Page, 2008. "agent-based models," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract. • Leigh Tesfatsion, 2006. "Agent-Based Computational Economics: A Constructive Approach to Economic Theory," ch. 16, Handbook of Computational Economics, v. 2, [pp. 831-880]. doi:10.1016/S1574-0021(05)02016-2. • Kenneth L. Judd, 2006. "Computationally Intensive Analyses in Economics," Handbook of Computational Economics, v. 2, ch. 17, pp. 881- 893. Pre-pub PDF. • L. Tesfatsion and K. Judd, ed., 2006. Handbook of Computational Economics, v. 2, Agent-Based Computational Economics, Elsevier. Description Archived 2012-03-06 at the Wayback Machine & and chapter-preview links. • Thomas J. Sargent, 1994. Bounded Rationality in Macroeconomics, Oxford. Description and chapter-preview 1st-page links.
- ^ • W. Brian Arthur, 1994. "Inductive Reasoning and Bounded Rationality," American Economic Review, 84(2), pp. 406-411 Archived 2013-05-21 at the Wayback Machine. • Leigh Tesfatsion, 2003. "Agent-based Computational Economics: Modeling Economies as Complex Adaptive Systems," Information Sciences, 149(4), pp. 262-268 Archived April 26, 2012, at the Wayback Machine. • _____, 2002. "Agent-Based Computational Economics: Growing Economies from the Bottom Up," Artificial Life, 8(1), pp.55-82. Abstract and pre-pub PDF Archived 2013-05-14 at the Wayback Machine.
- ^ • W. Brian Arthur, 1994. "Inductive Reasoning and Bounded Rationality," American Economic Review, 84(2), pp. 406-411 Archived 2013-05-21 at the Wayback Machine. • John H. Holland and John H. Miller (1991). "Artificial Adaptive Agents in Economic Theory," American Economic Review, 81(2), pp. 365-370 Archived 2011-01-05 at the Wayback Machine. • Thomas C. Schelling, 1978 [2006]. Micromotives and Macrobehavior, Norton. Description Archived 2017-11-02 at the Wayback Machine, preview. • Thomas J. Sargent, 1994. Bounded Rationality in Macroeconomics, Oxford. Description and chapter-preview 1st-page links.
- ^ • Joseph Y. Halpern, 2008. "computer science and game theory," The New Palgrave Dictionary of Economics, 2nd Edition. Abstract. • Yoav Shoham, 2008. "Computer Science and Game Theory," Communications of the ACM, 51(8), pp. 75-79 Archived 2012-04-26 at the Wayback Machine. • Alvin E. Roth, 2002. "The Economist as Engineer: Game Theory, Experimentation, and Computation as Tools for Design Economics," Econometrica, 70(4), pp. 1341–1378 Archived 2004-04-14 at the Wayback Machine.
- ^ Leigh Tesfatsion, 2006. "Agent-Based Computational Economics: A Constructive Approach to Economic Theory," ch. 16, Handbook of Computational Economics, v. 2, sect. 5, p. 865 [pp. 831-880]. doi:10.1016/S1574-0021(05)02016-2.
- ^ Demetrescu, Matei; Leppin, Julian S.; Reitz, Stefan (2020-06-11). "Homogeneous vs. heterogeneous transition functions in panel smooth transition regressions". Econometric Reviews. 40 (2): 177–196. doi:10.1080/07474938.2020.1773666. ISSN 0747-4938.
- ^ Farrell, Max H.; Liang, Tengyuan; Misra, Sanjog (2021). "Deep Neural Networks for Estimation and Inference". Econometrica. 89 (1): 181–213. doi:10.3982/ecta16901. ISSN 0012-9682.
- ^ "Deep learning for individual heterogeneity: an automatic inference framework". 2021-07-27.
{{cite journal}}: Cite journal requires|journal=(help) - ^ "The Impact of Machine Learning on Economics", The Economics of Artificial Intelligence, University of Chicago Press, pp. 507–552, 2019, retrieved 2022-05-05
- ^ Jesus, Browning, Martin Carro, (2006). Heterogeneity and microeconometrics modelling. CAM, Centre for Applied Microeconometrics. OCLC 1225293761.
{{cite book}}: CS1 maint: extra punctuation (link) CS1 maint: multiple names: authors list (link) - ^ Charpentier, Arthur; Élie, Romuald; Remlinger, Carl (2021-04-23). "Reinforcement Learning in Economics and Finance". Computational Economics. doi:10.1007/s10614-021-10119-4. ISSN 1572-9974.
- ^ "Oversight hearings on Public Law 93-410, the geothermal energy research, development and demonstration act of 1974: hearing before the Subcommittee on Energy Research, Development and Demonstration of the Committee on Science and Technology, U.S. House of Representatives, second session, January 20, 1976". 1976.
{{cite journal}}: Cite journal requires|journal=(help)