Feature Selection Library (MATLAB Toolbox)

TL;DR

This short report provides an overview of the feature selection algorithms included in the FSLib MATLAB toolbox among filter, embedded, and wrappers methods.

Abstract

The Feature Selection Library (FSLib) introduces a comprehensive suite of feature selection (FS) algorithms for MATLAB, aimed at improving machine learning and data mining tasks. FSLib encompasses filter, embedded, and wrapper methods to cater to diverse FS requirements. Filter methods focus on the inherent characteristics of features, embedded methods incorporate FS within model training, and wrapper methods assess features through model performance metrics. By enabling effective feature selection, FSLib addresses the curse of dimensionality, reduces computational load, and enhances model generalizability. The elimination of redundant features through FSLib streamlines the training process, improving efficiency and scalability. This facilitates faster model development and boosts key performance indicators such as accuracy, precision, and recall by focusing on vital features. Moreover, FSLib contributes to data interpretability by revealing important features, aiding in pattern recognition and understanding. Overall, FSLib provides a versatile framework that not only simplifies feature selection but also significantly benefits the machine learning and data mining ecosystem by offering a wide range of algorithms, reducing dimensionality, accelerating model training, improving model outcomes, and enhancing data insights.

Figures (1)

• Figure 1: (Left) In this example, features x and y are redundant, because feature x provides the same information as feature y with regard to discriminating the two clusters. (Right) In this example, we consider feature y to be irrelevant, because if we omit x, we have only one class, which is uninteresting.