Web Usage Mining Algorithm for an Academic Search Application

Web Usage Mining implies survey of information related to web log file. It can discover the browsing patterns and used for modeling the navigation behavior of web users. Analysis of log file gives an indication of the users' preferences. The data present in the log file cannot be used as it is, for the mining process. The contents of the log file should be preprocessed and put in a structured format. This paper focuses on design and execution of an algorithm for analyzing the Web Usage Log more specifically for an academic search application. It accepts the web log file and user queries as input and generates output containing the web log details as well as summary data. Descriptive Data Mining is considered for analyzing the web log file. The interestingness measures t-weight and d-weight, are computed. The model discussed in this paper will help significantly in improving user navigation, web design and usage patterns.