A comparison between extrinsic and intrinsic technique for multi-document text summarization

This paper presents, a comparison between Extrinsic and Intrinsic technique for multi-document text summarization has been presented. In extrinsic technique, lexical similarity measure and in intrinsic technique, semantic similarity measure between the sentences has been evaluated. The input data comprises of both the technical and literature based articles for in-depth assessment of the approach in both cases. Semantic or sense based similarity between a pair of sentences in intrinsic methodology has been obtained, by retrieving synonymous words from English WordNet-2.1. To conduct a comparative study, both the approaches have been tested on 5 sets of data comprised of at least 5 related documents in each set. Further, with the help of a linguistic expert, the results have been verified.