The Design And Realization Of A Personalized System Prototype Based On Tomcat Log Mining

In recent years, there have been a large number of domestic J2EE-based website systems going into service,however,the majority are seldom considered the problem that the information and contents of the websites are too much to be browsed by the users while the systems are being designed and developed. Meanwhile, a lot of log information created by Tomcat,which is a web servers implemented by J2EE standard can not be reasonably mined and used,and it also leads to information wasting .In order to merge these problems into one and solve it, after refering to the theory of Web Usage Mining, the paper treats the web log of the Tomcat as research object,then makes a systematic analysis and research on the process of Web Usage Mining,and integrates Web Usage Mining technology into the prototype of website system in its design and development stages to supply the personalized services for users and help them access to the useful web resource and information or improve the quality of the information obtained by them.This paper mainly focuses on the design and implementation of the personalized prototype system, and the following results have been achieved.After making the conclusion of the recent research in the web logs mining area both home and abroad, the paper discusses some related issues for the web log data pre-processing,moreover,a further and elaborate discuss or illustration in the five stages of the web log data pre-processing are given.They are data cleaning, user identification, session recognition, path completion,and frame page filtering. The paper gives the concrete plans for these five steps,especially,improves and optimized the methods for the realization of user identification and path completion,which have made data pre-processing more suitable for our personalized prototype system ,and simplified difficulties in order to enhance feasibility of our coding.On the basis of the web logs pre-processing research, the paper proposes an program for the web personalized system depending on Tomcat log mining. The program generally divides the system design into the total design and detailed design. The total design illustrates the layout and framework of the personalized system, according to the practical function ,the system is classified as two parts,which are on-line part and off-line part. As the core in this system, and also taking the different forms and function of the data through each task into account, the off-line part is resolved into the data preparation module ,which is responsible for the data collection and data pre-processing ; the data mining module ,which is responsible for the storage of mining algorithms of user model , data mining and the storage of the result; the visualization module,which is responsible for matching the individual information of each login user and displaying the comprehensible data forms; The detailed designs reasonedly illustrate the implement method of these three modules .Following the design in this paper, both the log pre-processing and mining algorithms have been firstly coded by using Java programming language . Secondly,the logs have been experimented and analysis results have been given in this paper.Thirdly,a Web News Information Management System has been realized by using JSP and AJAX technology. Finally, according to the overall design for the whole personalized system prototype in this paper, some code and achievement in log mining study have been integrated into the system seamlessly, to finish the task of prototype realization.And the performability and effectiveness of the program designed in this paper have been proved.