Graph Based Generation Of Research Article Summaries Using Definition Extraction

Abstract

newline With the explosion of information on web summary generation is essential for many applications Summaries are mainly useful for the research and learning community to get the overall gist of what is given in the document Most of the research articles do not contain summaries and normally learners skim through the abstract and conclusion to get an overall gist about the document Automatic summarization aids this by extracting the summary of a document This aids in representation of knowledge in a more useful and attractive manner This thesis proposes a systematic approach to generate research article summaries using concept graphs The semantic relationships between the sentences play a vital role in summary graph generation Semantic filters have been used to improve the quality of the graph being generated A taxonomy mapping is performed to capture the semantic relationship between sentences in research paper and to obtain a summarized graph.