Sequence Alignment Methods in Web Usage Mining

Abstract

The World Wide Web creates a universal space of information that can be accessed by individuals, companies, government, universities, students, teachers and business people. In this chapter, the World Wide Web is introduced by describing its origin, growth and how it became one of the world’s largest databases. Web Mining studies discover and analyze useful information from the World Wide Web. An overview of different topics in Web Mining studies is given. Finally, we describe and motivate our research statement, followed by the outline of this dissertation. … A well-known problem in studies of web usage behavior is extracting information about the order in which people visit web pages. If we are able to group temporary aspects of visiting patterns together based on how the order of pages occurs within patterns and if large groups are provided with similar order-based characteristics, we could use this information for the benefit of the visitor. For example, if 10% of the visiting patterns show that page x is visited before page y, and that page y is visited before page z, we may predict that page z will be visited after page x followed by page y. This way, a proxy server is able to provide faster deliveries of pages to web visitors. Also, if no direct hyperlinks exist from page x to y and from page y to z, we may suggest inserting direct hyperlinks for the ease and convenience of the visitor. Another advantage of using order-based information of groups of visiting patterns is suggesting cross-links between web pages. Cross-links are references to other documents based on common features. Especially for commercial web sites, cross-links may be an interesting asset. For example, people who visited product X are offered a cross-link to product Y. Through studying the order of visited pages we are able to examine which cross-links are used and which are not used. This means that, in the end, the results of cross-selling products X and Y through the web may be compared with the effect or use of cross-links. In other application domains like studies of the learning curve and psychology of the web visitor, the order of visited pages may provide information to construct guided tours for first time visitors and for visitors who frequently use the web site. ...