Optimal Features Set for Extractive Automatic Text Summarization

Abstract

The goal of text summarization is to reduce the size of the text while preserving its important information and overall meaning. With the availability of internet, data is growing leaps and bounds and it is practically impossible summarizing all this data manually. Automatic summarization can be classified as extractive and abstractive summarization. For abstractive summarization we need to understand the meaning of the text and then create a shorter version which best expresses the meaning, While in extractive summarization we select sentences from given data itself which contains maximum information and fuse those sentences to create an extractive summary. In this paper we tested all possible combinations of seven features and then reported the best one for particular document. We analyzed the results for all 10 documents taken from DUC 2002 dataset using ROUGE evaluation matrices.