Hierarchical Directed Acyclic Graph (HDAG) Based Preprocessing Technique for Session Construction

Web access log analysis is to examine the patterns of web site usage and the features of user's behavior. Preprocessing of the log data is very essential for efficient web usage mining as the normal log data is very noisy. Session construction is very vital step in the preprocessing phase and recently various real world problems can be modeled as traversals on graph and mining from these traversals provides effective results. On the other hand, the traversals on unweighted graph have been taken into consideration in existing works. This paper oversimplifies this to the case where vertices of graph are given weights to reflect their significance. Patterns are closed frequent Directed Acyclic Graphs with page browsing time. The proposed method constructs sessions using an efficient Directed Acyclic Graph approach which contains pages with calculated weights. Hierarchical Directed Acyclic Graph (HDAG) Kernel approach is used for session construction. The HDAG directly accepts several levels of both chunks and their relations, and then efficiently computes the weighed sum of the number of common attribute sequences of the HDAGs. This will help site administrators to find the interesting pages for users and to redesign their web pages. After weighting each page according to browsing time a DAG structure is constructed for each user session.