A new method for natural language call routing problem solving

Natural Language call routing remains a complex and challenging research area in machine intelligence and language understanding. This paper is in the area of classifying user utterances into different categories. The focus is on design of algorithm that combines supervised and unsupervised learning models in order to improve classification quality. We have shown that the proposed approach is able to outperform existing methods on a large dataset and do not require morphological and stop-word filtering. In this paper we present a new formula for term relevance estimation, which is a modification offuzzy rules relevance estimation for fuzzy classifier. We propose to split the classification task into two steps: 1) “garbage” class identification; 2) further classification into meaningful classes. The performance of the proposed algorithm is compared to several standard classification algorithms on the database without the “garbage” class and found to outperform them with the accuracy rate of 85,55 %. Combination of our approach with 9-NN algorithm for two-stage classification problem definition provides the accuracy rate of 77,11 % for test sample at whole.