Identification of User Behaviour by Web Usage Mining

Abstract

In the current scenario, information is multiplied everyday due to the technological development in World Wide Web and Internet. Millions of customers are daily accessing the Internet to search the information and fulfill their requirements. Information was overload in the web servers and receiving the right idea was a difficult task of the customers. So, the research in web mining is focused and key research area to facilitate the information searching process. Web mining has special place as it helps the organizations and website analyst to make clever decisions about their customers. Web usage mining collects the website visitor’s browsing details and recorded in the kind of plain text files in web servers. This information helps to identify the website visitor’s navigation behaviors. Web recommendation, one such methods of web personalization that endorses the web pages to its website visitors depends on its earlier browse history. Web recommendation schemes aid the website companies for better navigation of web pages, rapidly attain their terminus and to attain pertinent idea. In this research work an efficient web page recommendation technique is developed by using web usage mining and pattern mining algorithms to improve the accuracy of the existing web page recommendation techniques. The aim is to categorize user behavior in recognizing the designs of the browser and navigation data of web users as well as to evaluate the performance of the Frequent Pattern (FP) Growth algorithm, Apriori algorithm and modified web log searching algorithm. The association rule mining method was executed to its clustered web logs to identify the regularly visiting web pages through the users. Realtime datasets are evaluated. The proposed recommendation technique is compared against the existing recommendation techniques to prove that that the proposed recommendation technique improves the accuracy of the web page recommendation.