A clustering algorithm that groups users according to their Web access patterns was developed and its performance compared with the traditional k-means clustering algorithm. Results show that the new algorithm which is based on ART1 performs better in terms of intracluster distances. The technique was also applied in a prefetching scheme that predicts future user requests. Its prediction accuracy was as high as 97.78 percent.
ASJC Scopus subject areas
- Computer Science(all)