Automatic summarization of web documents

• Main Author: Jodihardja, Marcellus Reinaldo
• Other Authors: Mao Kezhi
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/54319
• Institution: Nanyang Technological University

Nowadays, we face an information overload, with all the rapid development in R& D and technological advancement. Even though information overload means that we can have various information regarding a specific topic, but it start to became more difficult to retrieve all the information needed in a limited time. The objective of this project is to create an auto-summarization program that can create a good summary of some documents in matter of seconds. By having this program, hopefully we can have all the information needed that are encapsulated in a dense and compact document. Latent Semantic Analysis is chosen to be the fundamental concept of this auto-summarization program. Thus, TFIDF (Term Frequency – Inverse Document Frequency) is utilized to give value of importance for each term, and Singular Value Decomposition is used to select the best sentences that can represent all information in a document. Some modifications have also been applied onto the algorithm in order to increase the efficiency and reduce the complexity time of this program. Furthermore, “meta” summarization method has also been implemented, to create a summary from some summaries that have been created from some input documents. This project successfully implemented all the algorithms needed and thus creating a good summary based on some input documents.