A Hybrid Approach for Automatic Extractive Summarization

Abstract

In recent times, there have been many works in automatic text summarization as it has become a very intriguing topic of natural language processing. A summary should be concise, delivering all the important facts of a document. State-of-the-art extractive text summarizers use sentence ranking in various ways to extract significant summary sentences. In this paper, a hybrid approach has been introduced for single document extractive summarization. A combination of approaches such as sentence-ranking based on key-phrases and sentiment analysis has been proposed. Moreover, the work combines another approach that picks summary sentences based on their interconnection with other sentences in the text for getting a better result. Through empirical experiments, the proposed approach has been found to generate better summaries than similar existing systems.