Ontology-based semantic smoothing model for biomedical document clustering. (2015)

• Record Type: Journal Article
• Authors: Logeswari, S.; Premalatha, K.
• Is Part Of: International journal of telemedicine and clinical practices. Volume 1:Number 1(2015)
• Page Start: 94
• Page End: 110
• Publication Date: 2015
• Subjects: semantic smoothing -- document clustering -- ontology -- Jaccard index -- silhouette index -- FM index -- biomedical documents -- data mining -- cluster quality -- domain ontology -- concept extraction -- text documents; Telecommunication in medicine -- Periodicals; Medical informatics -- Periodicals; 610.28
• Languages: English
• ISSNs: 2052-8434

Abstract

One of the major issues of data mining is the clustering of unstructured text documents. Traditional clustering algorithms are failing to prove the accuracy of the clustering process because of the characteristics of text documents such as high dimension, complex semantics, sparsity, etc. Recent researches focus on the clustering of text documents based on the semantic smoothing technique, which resolves the conflicts by general words and the sparsity of class-specific core words. In this work ontology-based semantic smoothing model is proposed which uses the domain ontology for concept extraction. It is a mixture of simple language model and a topic signature translation model. The results obtained from the proposed method shows a significant improvement in the clustering process than the existing methods in terms of cluster quality.