Supervised Feature Discretization with a Dynamic Bit-Allocation Strategy
Figueiredo, M. A. T.
Supervised Feature Discretization with a Dynamic Bit-Allocation Strategy , Proc Portuguese Conf. on Pattern Recognition - RecPad, Lisboa, Portugal, Vol. --, pp. -- - --, November, 2013.
Digital Object Identifier: 0
Download Full text PDF ( 151 KBs)
The use of feature discretization (FD) may be beneficial in several machine learning and pattern recognition tasks. By attaining adequate compact data representations, FD may improve the performance of many methods. It is often the case that learning with discrete data representations yields both lower training time and better accuracy, as compared to the original features. Moreover, FD may also allow for a better human
understanding/interpretation of the data. However, many FD techniques are sub-optimal, in the sense that they do not take into account feature interdependencies.
In this paper, we propose a dynamic supervised FD technique. Our method selects the discretization cut-points by simultaneously maximizing two criteria: the dependency between the discretized features and
the class label and the independence among these features. The method is built as an incremental bit allocation scheme, where mutual information (MI) is used as the dependency/independence measure that constitutes the underlying criterion. Experimental results on low and medium dimensional datasets show that the proposed method often achieves better accuracy than other well-known supervised FD approaches.