A New Fuzzy Hierarchical Classification Based on SVM for Text Categorization

  • Authors:
  • Taoufik Guernine;Kacem Zeroual

  • Affiliations:
  • Département d'Informatique, Université de Sherbrooke, Sherbrooke (Quebec), Canada J1K 2R1;Département d'Informatique, Université de Sherbrooke, Sherbrooke (Quebec), Canada J1K 2R1

  • Venue:
  • ICIAR '09 Proceedings of the 6th International Conference on Image Analysis and Recognition
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper we present a new fuzzy classification method based on Support Vector Machine (FHCSVM-Text) to treat categorization document problem. In context of document categorization, we have to separate large number of classes. SVM becomes an important machine learning tool to handle categorization document problem. Usually, SVM classifier is implemented to treat binary classification problem. In order to handle multi-class problems, we present a new method to build dynamically a fuzzy hierarchical structure from the training data. Our method consists in gathering the similar documents in the same class from the root until leaves, based on its textual content. The original problem is divided into sub-problems. The proposed method consists of three steps : (i) Preprocessing step to reduce the large number of features (ii) Fuzzy hierarchical classification (iii) and introducing SVM classifier at each node of the hierarchy. The fuzzy hierarchical structure extracts the fuzzy relationships between deferent classes. Our experimental results improve high accuracy in the Reuters corpus face standard document categorization techniques.