Preprocessing of Web Log Data for Web Personalization

World Wide Web has its impact on almost every facet of human lives. It is the prevailing and most popularly known information source that is very simply assessable and searchable. Prior to web, useful information can be gathered by referring a document or gathering the data from the expert's in the related areas, but with the rapid advancements in information technology, web has rigorously changed data seeking behavior of the users. The rapid growth in size and use of World Wide Web with its unique characteristics made web data mining an upcoming and area of demand in the present era. Web Usage Mining is the process of finding information from web usage logs that contain the click behavior of the user to attain the information the user needs. The procedure that is performed to improve the quality of raw data so as to improve the efficiency and ease of implementing pattern discovery methods is called data preprocessing. The web usage data that is taken from the server logs has to undergo many preprocessing phases so as to transform the data into a form that can be applicable to required mining task. In this paper we elucidate different preprocessing techniques applied to three different data sets.