Study Of Data Pre-processing Of The Personalized Service Of Economic And Trade Cooperation Between China And Russia

With the development of internet, Web is providing more and more information to us. However, the increase of Web site quantity and information explosion bring users more difficulties in looking for useful information. Accordingly, personalized service system emerges to satify the requirement of'one to one'sevices. The reality of personalized service relies on the technoligy of Web data mining. Generally, Web data mining can be devided to four task: search of recources, data pre-processing, pattern recognition and pattern analyse. Data pre- processing is one of the most important process of Web data mining. It also a most load-costing section. The quality of data pre-processing can directly infunce the result of data mining. For defferent fields, data pre-processing, pattern recognition and pattern analyse have their different processing content.Now, as far as the data pre-processing aiming at Personalized Service is concerned, most researches are only simply using heuristic formula to process log data, not combining with the relative domain knowledge. Especially in the field of session identification, people take the fixed value to judge on session's start and end point. The precision of session identification is poor, so that influence the data mining of users'usage hobby.This paper emphases on the research of data pre-processing course of Web data mining aiming at serving personalized service. The object of this research is creating data which can satisfy the command of Web data mining of personalized service. I build a Web log data pre-processing system by improving the heuristic formula, modeling, designing data structure and programming process. This system packages those functions of data pre-processing such as data cleaning, user identifying, session identifying, etc. The reality of this system can increase the reusing ability of data processing, making the researching of personalized service be more convenient. After executing this system, we can get a database in SQL SERVER which can provide data for the research of personalized service.We need to take more effort to research in these fields: how to process large scale web log data more efficiently, how to identify users and sessions more effectively, how to take advantage of the field knowledge in the data pre-processing course more sufficiently, etc.