Author Email: firstname.lastname@example.org
As we all know that, in today’s era World Wide Web places a measurable role by providing lots of information to the user which will be more useful. Thus, to know the process of discovering and analyzing usage patterns of user, the paper aims to predict the behavior of the user from the web logs. The weblogs is termed as web log mining or web usage mining, every click made by the user will be automatically entered into the weblogs of the corresponding web server. The proposed data structure, pattern tree can be used efficiently to store the usage patterns and their frequency of all the users using the path sharing. The Enhanced Pattern Tree (EPT) maintains the relationship between different patterns and the corresponding users as rules. Search for the specific pattern would yield the corresponding user and vice-versa in minimum no. of searches as rule sharing is used by the pattern tree data structure. This research work aims to develop a framework for analyzing user behaviour through user patterns obtained from the web server logs and supports the use of association rules for representing the relationship between user and patterns.
Today’s epoch World Wide Web plays a major role by providing a huge amount of information in all aspects so by this the user is very much familiar with the web data and the importance of the data for their daily needs, so this paper aims to predict the behavior of the user by providing the information related to the user based on their daily usage . Web mining is the use of data mining technique to automatically extract the information from the web data which consist of web documents, hyperlinks between documents and usage logs of websites, etc. There are three categories of web mining viz., Web usage mining, Web structure mining and Web content mining. These categories focus on knowledge discovery from the web. Web Usage Mining focuses on extracting useful information and patterns of the user for further use . The web usage mining mainly deals with the web server logs, which consists of information like IP address, client/user id, date, time, method, status code and size of the object. These log files are created and generated by the server and hence also referred to as server logs and are handled by the server administrator . By using these log files this paper aims to predict the user given the pattern and also pattern given the user. By processing these data, either using more complicated data mining techniques, or by using simple statistical methods, we can identify or predict the interesting area, and patterns concerning the activity in the Web site. This process includes 5 stages, namely Data Cleaning, User Identification, Session Identification, Enhanced Pattern Tree Construction , and Pattern Recognition . Initially, in data cleaning, erroneous data will be removed from the log file. Then each user is identified according to his/her IP address as specified in the log file which is known as user identification. In session identification, the time spent by each user on a particular website will be identified . The frequency of the users visiting a particular website will be represented in the form of a pattern tree in tree construction step. Finally, the usage pattern for every user is extracted. This process is performed by considering the frequency of users visiting each page . The knowledge thus extracted may be utilized to restructure the website so that efficient utilization of time is done. This work supports the use of association rules  for representing the relationship between user and patterns.
Web usage mining is indeed one of the emerging areas of research and important sub-domain of data mining and its techniques. Focus on the areas of preprocessing, including data cleaning, session identification, user identification. Used to analyze the user patterns from which information regarding the problems occurred to the users and usage of the website can be obtained within particular intervals of time.
463 total views, 1 views today