A NOVEL PREFETCHING TECHNIQUE THROUGH FREQUENT SEQUENTIAL PATTERNS FROM WEB USAGE DATA
Swamy T.M Veeragangadhara, Dr.G.T. Raju · COMPUSOFT: An International Journal of Advanced Computer Technology · 2015
Frequent sequential patterns (fsp) from web usage data (wud) are very important for analyzing and understanding users behavior to improve the quality of services offered by the world wide web(www). Web prefetching is one of the techniques for reducing the web latency there by improve the web retrieval process. This technique makes use of prefetching rules that are derived from fsps. In this paper, we explore the different fsp mining algorithms such as spm, fp growth, and spade for extraction of fsps from wud of an academic website for a period that varies from weekly to quarterly. Performance analysis on all of these fsp algorithms has been made against the number of fsps they generate with a given minimum support. Experimental results shows that spade fsp mining algorithm perform better compared to spm and fp growth algorithms. Based on the fsps, we propose a novel prefetching technique that generate prefetching rules from the fsps and prefetch the web pages so as to reduce the users’ perceived latency.