Document Clustering and Summarization Based on Association Rule Mining for Dynamic Environment

Document Summarization is a technique, which reduces the size of the documents and gives the outline and crisp information about the given group of documents. This paper introduces a new update summarization algorithm incorporating association rule mining and correlated concept based hierarchical clustering for dynamic environment. In this algorithm, the associated concepts are extracted using Rule mining technique (Generating Association Rules based on Weighting Scheme) and the Correlated concepts (terms and their related terms) are extracted based on concept extraction algorithm. Extracting concepts based on association rule, helps the user to cluster and summarize the similar concept, which in turn improves the quality of the cluster and the created summary. The performance of the hierarchical clustering based update summarization technique is compared with the existing COBWEB (update summarization) algorithm and static summarization algorithms namely; MEAD, CPLN (Centroid, Position, Length and Numerical value) and CPLNVN (Centroid, Position, Length Numerical value and VerbNoun) considering Precision, Recall and F-measure as performance metrics. Scientific literature and 20 Newsgroups are chosen as the data set for the experiment analysis. The experimental results demonstrate that the proposed algorithm exhibit better performance, compared to the existing algorithms for summarization.