A Novel Algorithm for Automatic Text Summarization System Using Lexical Chain

Abstract

In the field of text classification and information retrieval, the process of text summarization has always been an important aspect. It decreases the size of text and preserves its information content by providing a shorter illustration. This illustration has the major portion of data, i.e., the most vital information and it is no longer than the half of source data. In addition, it is the best solution for information overloading problem as we do not have to scan through each line of long length documents and still receive the foremost important information. For the betterment of this process, we have proposed an algorithm utilizing lexical chain calculation and it is implemented using Eclipse Java Development Tool, enterprise edition for web developers. Along with WordNet API, this method also included the nouns and proper nouns in the computation of lexical chains. MAX tagger is used for part-of-speech tagging and statistical calculations. By the execution of proposed methodology, it is clear that it is a better approach which resulted in better output in terms of (i) Execution time, as compared to the existing algorithm; (ii) Improved matching of words between the human-generated summary and proposed algorithm-generated summary; (iii) Better recall, that are commonly used criteria for summary evaluation.