SRL-GSM: a hybrid approach based on semantic role labeling and general statistic method for text summarization

Abstract

Sentence extraction techniques are commonly used to produce extraction summaries. The goal of text summarization based on extraction approach is to identify the most important set of sentences for the overall understanding of a given document. One of the methods to obtain suitable sentences is to assign some numerical measure of a sentence for summary called sentence weighting and then select the best ones. In this study, we propose Semantic Role Labeling (SRL) approach to improve the quality of the summary created by the general statistic method. We calculate a couple of sentence semantic similarity based on the similarity of the pair of words using WordNet thesaurus to discover the word relationship between sentences. We perform text summarization based on General Statistic Method (GSM) and then combine it with the SRL method. We compare our results with the baseline summarizer and Microsoft Word 2007 summarizers. The results show that SRL-GSM and GSM give the best average precision, recall and f-measure for creation of summaries.