An Approach for Frequent Access Pattern Identification in Web Usage Mining

Abstract

In the consent of this internet world, nobody is untouched with the internet for their usage. For such kind of scenario, data mining becomes an essential part of computer science. Data mining is a sub-field, which computationally processes the data, collected and is able to help the analysts of the research and development department, and the scientists, for proposing the ideas for some betterment of the organization. The user access is recorded in log files. The information regarding the website traversal of a user is always tracked by using these log files. These log files are generally stored at the server or at the client side. The web server logs provide important information. In the field of web mining the analysis of the web logs is done to identify the users' search patterns. In the existing approaches of finding the patterns, tree have been created which is based on the frequent access pattern identification. The creation of tree has increased the overhead of web usage. Therefore, Single Scan Pattern Algorithm has been proposed, which is based on use of database scan without creating any tree. The proposed algorithm would be able to increase the efficiency and decrease the overhead of unnecessary database scanning.