Exploiting Semantic Term Relations in Text Summarization

The traditional frequency based approach to creating multi-document extractive summary ranks sentences on scores computed by summing up TF*IDF weights of words contained in the sentences. In this approach, TF or term is calculated how frequently a (word) occurs input and way does not take into account semantic relations among terms. paper, we propose methods that exploits for improving sentence ranking redundancy removal steps summarization system. Our proposed system has been tested DUC 2003 2004 benchmark datasets. experimental results reveal performance our text summarizer significantly improved when distributional similarity measure used finding relations. also outperforms some well known baselines which it compared.