FEATURE BASED EXTRACTIVE SUMMARIZATION (FBES) TECHNIQUE FOR LONG TEXT DOCUMENTS

Abstract

Text summarization is a sensible area of research, owing to the increased utilization of text data. The process of presenting the content in a condensed form while maintaining the data's integrity is summarization. The development and expansion of automatic text summarization techniques are remarkable these days, due to their widespread applicability. The lengthy information must be processed such that the text summary is created. The text summarising research problem has been addressed in a large number of published studies, but there are not many solutions for processing lengthy documents. The pre-processing, the computation of inter-sentence similarity, and the generation of the summary are the three key phases that are discussed in this research article, which presents the FBES technique for the summarization of lengthy documents. The inter-sentence similarity is computed by the degree of information and ROUGE score of sentences. The generated summaries are evaluated against the gold summary and the results prove that the generated summary shows better ROUGE-N, cohesion, sensitivity and readability scores. The average best ROUGE-1 and ROUGE-2 scores of the proposed work are 44.71 and 18.56 respectively, these scores are roughly 3% higher than those received by the present method.