Genetic algorithm based sentence extraction for text summarization

Abstract

The goal of text summarization is to generate summary of the original text that helps the user to quickly understand large volumes of information available in that text. This paper focuses on text summarization based on sentence extraction. One of the methods to obtain suitable sentences is to assign some numerical measure for sentences called sentence weighting and then select the best ones. The first step in summarization by extraction is the identification of important features. In this paper, we consider the effectiveness of the features selected using Genetic Algorithm (GA). GA is used for the training of 100 documents in DUC 2002 data set to learn the weight of each feature, which is evaluated using recall measurement generated by ROUGE for a fitness function. The weights obtained by GA were used to adjust the important features score. We compare our results with Microsoft Word 2007 summarizer and Copernic summarizer both for 100 documents and 62 unseen documents. The results show that the best average precision, recall, and f-measure for the summaries were obtained by GA.Â.