Medical text categorization using SEBLA and Kernel Discriminant Analysis

Education data mining (EDM) is an emerging field which take advantage of natural language processing, data mining, statistics and machine learning algorithms for different types of educational data especially related to medical e-learning. Medical text categorization is the one of the major component to help students more easily and effectively search medical text for e-Learning. In this paper, spectral based Kernel Discriminant Analysis has been introduced for medical text categorization. We evaluated the proposed approach on 10 most frequent categories of cardiovascular diseases group from Ohsumed data sets. When compared with existing approaches, the results have indicated significant increase in performance. In order to further refine the search, Semantic Engine that uses Brain-Like approach (SEBLA) is also introduced in this paper.