S-estimator
The goal of S-estimators is to have a simple high-breakdown regression estimator, which share the flexibility and nice asymptotic properties of M-estimators. The name "S-estimators" was chosen as they are based on estimators of scale.
We will consider estimators of scale defined by a function , which satisfy
- R1 – is symmetric, continuously differentiable and .
- R2 – there exists such that is strictly increasing on
For any sample of real numbers, we define the scale estimate as the solution of
,
where is the expectation value of for a standard normal distribution. (If there are more solutions to the above equation, then we take the one with the smallest solution for s; if there is no solution, then we put .)
Definition:
Let be a sample of regression data with p-dimensional . For each vector , we obtain residuals by solving the equation of scale above, where satisfy R1 and R2. The S-estimator is defined by
and the final scale estimator is then
.[1]
References
[edit]- ^ P. Rousseeuw and V. Yohai, Robust Regression by Means of S-estimators, from the book: Robust and nonlinear time series analysis, pages 256–272, 1984