Information Gain Based Term Weighting Method for Multi-label Text Classification Task

Abstract

In text classification, terms are given weights using Term Weighting Scheme (TWS) in order to improve classification performance. Multi-label classification task are generaly simplified into several single-label binary task. Thus, the term distribution are considered only in terms of positive and negative categories. In this paper, we propose a new TWS based on the information gain measure for multi-label classification task. This TWS try to overcome this shortness without affecting the complexity of the problem. In this paper, we examine our proposed TWS with eight well-known TWS on two popular problems using 5 learning algorithms. From our experimental results, our new proposed method outperforms other methods specialy regarding the macro-averaging measure.

Two new feature selection metrics for text classification

“…In literature, there are many feature selection metrics. Some of them are used term weighting purpose like IG and CHI, some of them are used for feature selection like Term Frequency-Inverse Document Frequency (TF-IDF) originated from the term weighting method [10][11][12][13]. Relevance Frequency (RF) is another term weighting metric.…”

“…In Equation ( 12), precision value (π ) and in Equation ( 13), recall value (ρ) are given. π = TP TP + FP (12) ρ = TP TP + FN (13) F-measure is given in Equation (14).…”

Generating Term Weighting Schemes Through Genetic Programming

“…Unsupervised TWSs are generally borrowed from Information Retrieval domain [26] and adopted for TC [22,7,23].…”

“…Several comparative studies on these TWSs for both term-weighting and feature selection has been reported in [31,24,8,22]. A new approach for term-weighting based on (TF-IG) have been proposed for multi-labeled classification task in [23]. The method computes a score based on all categories and then subtracts it from the original TF-IG weight.…”