Comparative Analysis of Automatic Text Summarizers

Abstract

In view of the intense time crunch faced by humans in today’s time, the idea of information being presented to people in the most concise and crisp manner possible is pretty much a necessity. This project entitled “COMPARATIVE ANALYSIS OF AUTOMATIC TEXT SUMMARIZERS” aims to summarize text documents automatically using software drawing from the concepts of machine learning and neural networks. There are two major kinds of Automatic Text Summarization (ATS) techniques, Extractive Text Summarization and Abstractive Text Summarization. Extractive ATS refers to picking out more important sentences, ones that more central to the chief idea of the document and stringing them together to give a summarised document. Abstractive ATS, on the other hand, deals with rephrasing the sentences in the original document and expressing them differently. These algorithms are generally more difficult to implement as this challenges the system to strain itself with complex skills such as paraphrasing and abstraction. In this project, Extractive ATS is done using lexical analysis. The results and a detailed analysis has been presented in this report. The project will later use machine learning algorithms to achieve ATS both extractive and abstractive in nature along with presenting a structured study of the difference in efficiencies among the various algorithms. I will also integrate the project to a front-end user interface that shall be user friendly and responsive.