Naive Bayes Classification for Subset Selection

TL;DR

This paper introduces the bx algorithm, extending Naive Bayes to multi-label classification for predicting unordered, interdependent item subsets from features, with demonstrated effectiveness on real-world data.

Contribution

The paper presents a novel bx algorithm that adapts Naive Bayes for multi-label subset selection, addressing interdependence among labels.

Findings

Effective on real-world multi-label datasets

Handles interdependent labels in subset prediction

Extends Naive Bayes to multi-label domain

Abstract

This article focuses on the question of learning how to automatically select a subset of items among a bigger set. We introduce a methodology for the inference of ensembles of discrete values, based on the Naive Bayes assumption. Our motivation stems from practical use cases where one wishes to predict an unordered set of (possibly interdependent) values from a set of observed features. This problem can be considered in the context of Multi-label Classification (MLC) where such values are seen as labels associated to continuous or discrete features. We introduce the \nbx algorithm, an extension of Naive Bayes classification into the multi-label domain, discuss its properties and evaluate our approach on real-world problems.