Nyaya based Automatic Quality Value Classifier for Concept Enrichment Behera Niyati Kumari*, Mahalakshmi G. S.** *Research Scholar, Department of CSE, CEG, Anna University, Chennai, Tamilnadu, India **Department of CSE, CEG, Anna University, Chennai, Tamilnadu, India Online published on 23 March, 2017. Abstract Concept descriptors like attributes, part-relations or semantic role mined from unstructured text can add a different dimension to the Morden knowledge representation models. Though quite a lot of works have been done for NON-QUALITATIVE attributes like part-of-relation extraction, comparatively little attempts has been made to extract QUALITATIVE attributes (i.e. size, colour, taste etc) from text. Taking this as the motivation, in this paper we propose an automatic, unsupervised and domain independent methodology to retrieve these “attributes” also known as “qualities” from unstructured text. These qualities are the basic characteristics like taste, colour, dimension etc. of any object or concept. This unique approach is based on Indian Logic system, Nyaya Sastra, an ancient classification framework for real world entities. The proposed framework includes two phases- (i) acquisition of quality values from unstructured text (ii) classifying the values according to Nyaya classification scheme. We have compared the performance of our system with the results obtained from a supervised a supervised approach. Ours system is the first of its kind to classify quality values into their respective quality categories. We have performed extensive analysis and experiments to show the reliability of our system. Top |