Text Summarization System: An Extractive Approach using Hierarchical Text Clustering

Abstract

The need for summarizing texts evolves from the large amount of data present in electronic channels which leads to distraction of users and wastage of their time.There are generally two major techniques for text summarization: extractive method and abstractive method.The extractive method has proven to be quite reliable and involves extracting the key sentences from the document to form a summary.In this paper, an unsupervised text mining model is developed for clustering and summarizing texts.The model is deployed into a web-based system for summarizing large documents.Using the informational criteria of redundancy, coherence, speed and information coverage, our approach chooses 'not likely', 'high', 'fast', and 'medium' as semantic dimensions values for the criteria respectively.