Analytic Feature Selection for Support Vector Machines

TL;DR

This paper introduces a novel filter-based feature selection algorithm for linear SVMs that leverages geometric properties of feature sets, showing promising results on high-dimensional text data.

Contribution

It develops an analytic feature selection method based on geometric properties, moving beyond heuristic approaches for linear SVMs.

Findings

Achieves excellent results on high-dimensional text datasets.

Identifies six geometric properties correlated with classifier performance.

Organizes features into meaningful types like unigrams and bigrams.

Abstract

Support vector machines (SVMs) rely on the inherent geometry of a data set to classify training data. Because of this, we believe SVMs are an excellent candidate to guide the development of an analytic feature selection algorithm, as opposed to the more commonly used heuristic methods. We propose a filter-based feature selection algorithm based on the inherent geometry of a feature set. Through observation, we identified six geometric properties that differ between optimal and suboptimal feature sets, and have statistically significant correlations to classifier performance. Our algorithm is based on logistic and linear regression models using these six geometric properties as predictor variables. The proposed algorithm achieves excellent results on high dimensional text data sets, with features that can be organized into a handful of feature types; for example, unigrams, bigrams or semantic structural features. We believe this algorithm is a novel and effective approach to solving the feature selection problem for linear SVMs.