CLUSTERING BASED DOCUMENT SUMMARIZATION

Abstract: Document summarization involves summarizing document as the information is continuously increasing with such a huge amount. Users do not have much time to spend reading thousands of lines. Today users want maximum information which describes everything and occupies minimum space. This paper discusses an improved approach for document summarization by using clustering. Summarization is process of producing single summaries from a document. The three major problems that were introduced in single document summarization were coped in k means clustering summarization i.e. coping with redundancy, coherency in summary, identifying difference in sentences. To identify similarity in documents various similarity measures are used i.e. similarity between the sentences of documents and then grouping them in clusters based on their tf*idf values of the words.