Publication

Abstract : Most of the text retrieval and mining methods are still based on the exact word matching and they use term frequency (word or phrases) as basic measure. It captures the importance of the term in the document but may not capture the original semantics of the term, resulting in poor retrieval performance. To overcome the lack of semantic consideration, a new framework has been introduced which relies on the concept based mining model and semantic based approach. The core part of our model is concept extraction, which perform functions such as document cleaning, parts-of-speech tagging, parsing, term and phrase extraction, feasibility analysis and relation miner. Semantic net and synonym dictionary preserve the semantic relationship in the text document. The dataset used here is ACM abstract articles collected from ACM digital library. Large sets of experiments using the proposed model were conducted and the results demonstrate the accuracy of mining model using semantics preserved concepts, feasibility analysis using singular value decomposition and semantic net representation.