The Research And Application Of Web-based Data Mining In Business Web Site

With the quick development of the Internet，World Wide Web brings us massinformation resources. Modern enterprises grow rapidly. In the process of informationaccumulated more useful data and useless one. Such countless data brings a certainamount of troubles. And how to obtain the useful information becomes a confusingproblem. The effective way to solve this problem is to apply traditional data miningtechnology on network. As the major part of data mining technique, the application ofweb data mining using on internet marketing and customer message analysis is veryimportant.With the significantly increased of data, it brings more opportunities and troublesfor the internet companies. The website construction and maintenance is very important.For large portals or e-commerce site, how to manage the successful operation becomes avery important one of these companies work. Successful or not, it needs not only thenecessary financial support to improve beyond the initial construction, but also needs topay attention to the latter part of the update and maintenance. A web site will be morevitality and more competitive with a good operation method. It will be unbeaten andbenefit from the crisis. In the process of a web site operation, it involves a lot of content,such as promoting corporate culture, website marketing management, websitemaintenance and updates later, the internal relations operation, etc. These managementworks decide the site’s business and future direction. The effectiveness of supportivedecision-making is a key factor.Decision support requires a lot of useful data. The date comes from the operationof the site itself. we must use an effective method-data mining technique to get theuseful information. Now, the data mining technology is flourishing, and is widely used.The applications of data mining technique focus on telecommunications, insurance,e-commerce and commercial chains, banks and some government departments. Forexample, sales records in a mall can dig out the specific customer’s favorite productportfolio, and find that some of the potential loss of customers in buying and customer characteristics, or some favorite customers of new products, etc. It shows that datamining technology can really determine the success of a business enterprise.This article will do some constructive research as the follows.First, this paper introduces the basic theory of internet marketing, describes themanagement models and does some necessary analysis on the encountered operationalissues. Second, the basic principle, the process and the roles of data mining aresummarized. Comparing with several decision support tools, this article does somespecial analysis on the characteristics of the enterprise’s internet marketing data. Then,it will extract the web site information and do some individual research. As theminimum unit, the web block is used in the form of web information analysis and thedesign of the web information extraction algorithms. Finally, based on the consumer’sindividual character differences and the cooperation’s overall marketing strategy, it willdesign the web data mining model. This model is divided into three parts, web side datacollection, and log preprocessing and algorithm analysis. This model is used on the datamining of a tourism website. It proves that this is a correct and effective method.This article is focus on e-commerce website in different categories of data miningand information web data extraction algorithm, especially the extraction research ine-commerce website theme keyword data. It raises some solution methods for extractingrelevant data from the site. In some extent, it becomes more efficiency and accuracy fordata mining. In the course of the study, it finds that in the past there are some problemsin information extraction technology. Such as regarding the whole web page as thesmallest unit of data extraction method, the accuracy is very poor. It needs for a smallerdivision to meet the web site needs. Based on this, the author does a further division forthe whole web page. The total area will be divided into different extraction blocks. Thedata extracting for different regions will be done individually. Considered the differentfocus, new threshold valve is endowed to each region. So the data extraction work iswell done.Clustering methods is used in this article for the web site log data mining. It mayextract some implicit information. Specific approach is that putting the high similarityof customer interest into a class. The clustering process can also be done on-line oroff-line. It can greatly reduce the level of algorithm time and improve the efficiency ofweb analysis by this method. And it will do well in solving the data sparseness problem,making the data more reasonable after the extraction. From the analysis of experimental results, this paper does some necessarysummarize about the data extraction and data mining algorithms. So, we get somerevelations that drawing the opening logical framework of Internet marketing strategy.It will enhance the customer’s loyalty and get the exact locations for the enterpriseproducts.Data mining is finally done on a travel web site by using this model. It shows thatthis model is feasible and the algorithm is proved to be reasonable and effective.In this research, due to personal knowledge structure, the time, and many other factors,it will inevitably bring out some questions:1、The coordination between efficiency and accuracy: Efficiency and accuracy is acontradiction in data mining process. Under the premise of guaranteeing the dataextraction speed, how to improve the accuracy of data extraction needs further study.2、 Customer privacy information protection: It will inevitably violate some secretinformation in customer preference data extraction process. Under the premise ofensuring the effectiveness of decision-making, how to make reasonable protection of thepersonal individual information is worth doing further discussions.