KOMBINASI WEB USAGE MINING DAN WEB CONTENT MINING UNTUK MENGHASILKAN PROFIL NAVIGASI PENGUNJUNG WEBSITE; COMBINATION OF WEB USAGE MINING AND WEB CONTENT MINING FOR GENERATING WEBSITE USER NAVIGATION PROFILE

HERWANTO, GUNTUR BUDI, N

2016 | Disertasi | FMIPA

The World Wide Web is now the primary source for information discovery. A user visit websites that provide information and browse the particular information in accordance to their topic interest. All the activities of the visitors recorded in an access log. Access logs can be an important source for finding useful navigational patterns and visitor behaviour. Website content can also be an important source to see the overall semantic structure of the website’s content. Both of these data sources can be combined to obtain an interesting navigational pattern of overall visitors. In this research, Latent Dirichlet Allocation (LDA) used to get the semantic structure of the entire content of the website. Results from LDA combined with a session that has been formed from the access log. To group the same navigation pattern, clustering algorithm, Self Organizing Map is used. Navigational patterns that has been grouped within each cluster become a source for sequential pattern mining algorithm, Prefixspan, to get some relevant navigational patterns. Evaluation shows that system can make a navigational profile that match overall navigational profile for new users.

Kata Kunci : N