Performance enhancement in session identification

Web logs are user transactions stored in the web log files, and it is unstructured for mining knowledge. Since the usage of internet has increased, data in the web log file is very important for finding user navigational behavior. Accurate web log mining extracts results and efficient navigational patterns that are crucial for tuning websites and consequently helping in visitor retention. Web log mining task is to extract some hidden knowledge which is not done by any other conventional data mining algorithms. In order to produce good result web log mining depends on some quality input. Hence the first challenge is to perform effective data cleaning and preprocessing, and the second challenge is to improve the efficiency of online prediction. This paper proposes an enhancement to web log mining and to evaluate its performance to fulfill the aforementioned challenges.