Using Case-Based Reasoning Approach for Text Classification Task

Abstract

This paper discusses the method of solving the text classification task by using case-based reasoning. As a part of the text classification module in the implemented system, modified TF-IDF (Term Frequency - Inverse Document Frequency) measure of the text content based on the known information on the distribution of documents between different categories is proposed. This modification allows to improve the quality of the classification by considering the information on the distribution of the words in the entire case base. Implementation of the prototype of text classification module used in the interactive voice response system is presented. This module allows to get away from button menus to automated voice interaction between the user and the system. The result of computational experiments confirming the effectiveness of the developed system is proposed.