Learning for semantic classification of conceptual terms
Document Type
Conference Proceeding
Publication Date
12-1-2007
Abstract
Extraction of concepts and identification of their semantic classes are useful in applications such as automatic instantiation of ontologies and construction of information extraction systems. Even though various techniques exist for the extraction of domain specific concepts from unstructured texts, very little concentration is in the semantic class labeling for concepts. In this paper we propose the Semantic Class Labeling(SCL) problem and differentiate it from the Named Entity Classification(NEC) problem. We also present a Naive Bayes solution to SCL. Experiments suggest that Naive Bayes learning method with specified features achieves high classification accuracy. Empirical and statistical evaluation on the significance of attributes for SCL is also presented. © 2007 IEEE.
Publication Source (Journal or Book title)
Proceedings - 2007 IEEE International Conference on Granular Computing, GrC 2007
First Page
253
Last Page
258
Recommended Citation
Punuru, J., & Chen, J. (2007). Learning for semantic classification of conceptual terms. Proceedings - 2007 IEEE International Conference on Granular Computing, GrC 2007, 253-258. https://doi.org/10.1109/GRC.2007.4403105