Abstract

Most of the existing web sites organize their content in a hierarchical manner. This organization may not be clear to the visitors because each visitor may have their own expected organization. For instance, it is often unclear to a visitor where a specific document is located. Usage knowledge, discovered from web usage mining, on the way visitors navigate in a web site could prevent disorientation, help the web site owner in designing the web site, provide efficient access between highly correlated object, and make better marketing decisions such as putting advertisements in proper places. In this report, we will give an overview on the web usage mining process with special emphasis on presenting two data mining techniques: association rules and path traversal pattern discovery. We introduced the notion of context awareness web usage mining which is a constraint pushed into these data mining techniques. As well, we will present our design and implementation of a web usage mining system call WUM. Finally, we will show the experimentation of using our WUM to mine for usage knowledge for a Computer Science department's web site.