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

Ahmad Mazyad (1) , Fabien Teytaud (1) , Cyril Fonlupt (1)

1 LISIC - Laboratoire d'Informatique Signal et Image de la Côte d'Opale

In text classification, terms are given weights using Term Weighting Scheme (TWS) to improve classification performance. Multi-label classification task is generally simplified into several single-label binary tasks. Thus, the term distribution is considered only in terms of positive and negative categories. In this paper, we propose a new TWS based on the information gain measure for the 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 five learning algorithms. From our experimental results, our new proposed method outperforms other methods, especially regarding the macro-averaging measure.

Domaines: Intelligence artificielle [cs.AI]; Machine Learning [stat.ML]

Ahmad Mazyad, Fabien Teytaud, Cyril Fonlupt. Information Gain Based Term Weighting Method for Multi-label Text Classification Task. Intelligent Systems Conference (IntelliSys) 2018, Sep 2018, London, United Kingdom. ⟨hal-01859697⟩