User:Bassis/Complexity Indices

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Besides complexity intended as a difficulty to compute a function (see computational complexity), in modern Computer Science and in Statistics another complexity index of a function stands for denoting its information content, in turn affecting the difficulty of learning the function from examples. Complexity indices in this sense characterize the entire class of functions to which the one we are interested in belongs. Focusing on Boolean functions, the detail of a class of Boolean functions c essentially denotes how deeply the class is articulated.

To identify this index we must first define a sentry function of . Let us focus for a moment on a single function c, call it a concept defined on a set of elements that we may figure as points in an Euclidean space. In this framework, the above function associates to c a set of points that, since are defined to be external to the concept, prevent it from expanding into another function of . We may dually define these points in terms of sentinelling a given concept c from being fully enclosed (invaded) by another concept within the class. Therefore we call these points either sentinels or sentry points; they are assigned by the sentry function to each concept of in such a way that:

  1. the sentry points are external to the concept c to be sentineled and internal to at least one other including it,
  2. each concept including c has at least one of the sentry points of c either in the gap between c and , or outside and distinct from the sentry points of , and
  3. they constitute a minimal set with these properties.

The technical definition coming from (Apolloni 2006) is rooted in the inclusion of an augmented concept made up of c plus its sentry points by another in the same class.

Definition of sentry function[edit]

For a concept class on a space , a sentry function is a total function satisfying the following conditions:

  1. Sentinels are outside the sentineled concept ( for all ).
  2. Sentinels are inside the invading concept (Having introduced the sets , an invading concept is such that and . Denoting the set of concepts invading c, we must have that if , then ).
  3. is a minimal set with the above properties (No exists satisfying (1) and (2) and having the property that for every ).
  4. Sentinels are honest guardians. It may be that but so that . This however must be a consequence of the fact that all points of are involved in really sentineling c against other concepts in and not just in avoiding inclusion of by . Thus if we remove remains unchanged (Whenever and are such that and , then the restriction of to is a sentry function on this set).

is the frontier of c upon .

A schematic outlook of outer sentineling functionality

With reference to the picture on the right, is a candidate frontier of against . All points are in the gap between a and . They avoid inclusion of in , provided that these points are not used by the latter for sentineling itself against other concepts. Vice versa we expect that uses and as its own sentinels, uses and and uses and analogously. Point is not allowed as a sentry point since, like any diplomatic seat, it should be located outside all other concepts just to insure that it is not occupied in case of invasion by .

Definition of detail[edit]

The frontier size of the most expensive concept to be sentineled with the least efficient sentineling function, i.e. the quantity


is called detail of . spans also over sentry functions on subsets of sentineling in this case the intersections of the concepts with these subsets. Actually, proper subsets of may host sentineling tasks that prove harder than those emerging with itself.

The detail is a complexity measure of concept classes dual to the VC dimension . The former uses points to separate sets of concepts, the latter concepts for partitioning sets of points. In particular the following inequality holds (Apolloni 1997)

See also Rademacher complexity for a recently introduced class complexity index.

Example: continuous spaces[edit]

Class C of circles in has detail , as shown in the picture on left below. Similarly, for the class of segments on , as shown in the picture on right.

Two points outside c (thick circle) are sufficient to prevent a larger circle not containing them from including it
The class of segments in and two points needed to sentinel its concepts

Example: discrete spaces[edit]

The class on whose concepts are illustrated in the following scheme, where “” denotes an element belonging to , “” an element outside and a sentry point:

This class has . As usual we may have different sentineling functions. A worst case , as illustrated, is: . However a cheaper one is :


  • Apolloni, B (2006). Algorithmic Inference in Machine Learning. International Series on Advanced Intelligence. 5 (2nd ed.). Adelaide: Magill. Advanced Knowledge International  Unknown parameter |coauthors= ignored (|author= suggested) (help)
  • Apolloni, B. (1997). "PAC learning of concept classes through the boundaries of their items". Theoretical Computer Science. 172: 91–120.  Unknown parameter |coauthors= ignored (|author= suggested) (help)