Abstract

In this age of Internet, Natural Language Processing (NLP) techniques are the key sources for providing information required by users. However, with the extensive usage of available data, a secondary level of wrappers that interact with NLP tools have become necessary. These tools must extract a concise summary from the primary data set retrieved. The main reason for using text summarization techniques is to obtain this secondary level of information. Text summarization using NLP techniques is an interesting area of research with various implications for information retrieval.

This report deals with the use of Latent Semantic Analysis (LSA) for generic text summarization and compares it with other models available. It proposes text summarization using LDS in conjunction with open-source NLP frameworks such as Mahout and Lucene. The LSA algorithm can be scaled to multiple large-sized documents using these framworks. The performance of this algorithm is then compared with other models commonly used for summarization and Recall-Oriented Understudy of Gisting Evaluation (ROUGE) scores. This project implements a text summarization framework, which uses available open-source tools and cloud resources to summarize documents from many languages such as, in the case of this study, English and Hindi.

Publication Date

6-2014

Document Type

Thesis

Student Type

Graduate

Degree Name

Computer Science (MS)

Department, Program, or Center

Computer Science (GCCIS)

Advisor

Rajendra K.Raj

Advisor/Committee Member

Xumin Liu

Advisor/Committee Member

Leon Reznik

Campus

RIT – Main Campus

Plan Codes

COMPSCI-MS

Share

COinS