Text Summarization by Hybridization of Hypergraphs and Hill Climbing Technique

Abstract

Automatic text summarization (ATS) is an application of natural language processing (NLP). It is the process of compressing the given text to create a summary. The challenge is creating a concise, non-redundant, coherent and inclusive summary that features all the significant points of the given text. The two approaches of summarization are extractive and abstractive. Extractive summarization works by choosing important sentences of the original text. It relies on the statistical relationship between the sentences. Since the sentences of the text are related by n-ary relationships, the authors have used these relationships to constitute the hyperedges of a hypergraph, which is called the sentence hypergraph. Hill climbing is an optimization technique that the authors have chosen to construct the sentence hypergraph. They succeeded in using the Helly property to select the significant sentences as summary. They have evaluated the performance of the system against the Gold summary using the ROUGE evaluation system.