Home // ICIW 2011, The Sixth International Conference on Internet and Web Applications and Services // View article

Decision Method of Training Data for Web Prefetching

Authors:
Zhijie Ban
Feilong Bao

Keywords: web prefetching; sliding window; training data

Abstract:
Web prefetching is an effective technique to reduce user-perceived latency. Most studies mainly focus on prediction algorithm but they ignore selection strategy of training data which is an important part of web prefetching. This paper presents a decision method based on monitoring prediction precision. It divides user access sequence into different data blocks and the changing features of prediction precision among data blocks indicate whether some training data is outdated. According to the varying trend of prediction precision, some user access requests are inserted into or deleted from training data. We use two real web logs to examine this proposed method and the simulation shows that our method can significantly improve prefetching performance.

Pages: 104 to 111

Copyright: Copyright (c) IARIA, 2011

Publication date: March 20, 2011

Published in: conference

ISSN: 2308-3972

ISBN: 978-1-61208-124-3

Location: St. Maarten, The Netherlands Antilles

Dates: from March 20, 2011 to March 25, 2011