Multi-Document Summarization Techniques: A Comparative Study

Multi-Document Summarization Techniques: A Comparative Study

Mr. Saish Hanuman Mahale

Masters of Engineering,

Department of Information Technology, Goa College of Engineering, Ponda, Goa 403401

saishmahale4165@gmail.com

Mr. Amogh Sanzgiri

Assistant Professor,

Department of Information Technology, Goa College of Engineering, Ponda, Goa 403401

amoghs@gec.ac.in

Keywords: Text Summarization, Multi Document Summarization, TextRank,  LexRank, Centroid Based Summarization, Maximal Marginal Relevance, K-Means Clustering.

Abstract— Text summarization is the process of generating a shorter version of a longer text while retaining its most important information. The aim of text summarization is to extract the most important information from a document and present it in a concise and easy-to-read format. There are two main types of text summarization: extractive and abstractive. Multi-document summarization is a text summarization technique that aims to generate a summary from multiple documents instead of just one. There are various algorithms for implementing Multi-document summarization such as TextRank, LexRank, Centroid Based Summarization,  Maximal Marginal Relevance, Cluster-based Sentence Selection (CBSS), SumBasic and K-Means Clustering. In this paper, we present an overview of Multi-document summarization and five methods: TextRank, LexRank, Centroid Based Summarization,  Maximal Marginal Relevance and K-Means Clustering. Also we have discussed how these algorithms can be evaluated using several metrics such as Precision, Recall, F1 Score, Pyramid Score, Content Selection Score, Sentence Compression Ratio.