Web Log Data Mining Based On Intelligent Computation

The ability of information production and data collection has been strengthened with the progressing networking and database technology. Lots of databases are used in the commercial management, the administrative work, the scientific research and the project development. But the magnanimous data has initiated the new question. These are the extremely strong message resources. The mechanism of tradition data retrieval and the method of statistical analysis cannot meet the needs of the information withdraws effectively. It only can become the"cloth wrapper"if the resources cannot serves for the enterprise and strategic development. Now the data mining technology appeared and it can discover the knowledge from the database.Data Mining is the process of withdrawing knowledge and information which are unknown, latent and useful from the data of massive, not incomplete and stochastic. Data mining technology includes characteristic, classification, connection, cluster, deviation, time series, and trend analysis and so on.The data is the non-structure, dynamic although there has lots of data on Internet and the complexity of the web page beyond the text. It just like look for a needle in a haystack which people want to find the data which oneself wants to. The superintendent of site need the automatic assistance designs tool to adjust page structure, improve service and develop electronic commerce for meeting the visitor's need according to user's visit interest, visit frequency, visit time. The visitors hope to look the individuality page and obtain the better service. The method of resolving the aspects is to use the data mining based web. It means to use the thought and method of data mining on the web to discover the useful information. The objects of data mining are not only the relational database, but also the valuable information.Web data mining is mainly divided into three components: Web content mining, Web structure mining and Web usage mining. Web usage mining is the chief study of this thesis. Web usage mining can find users'access patterns through mining Web server log. It is also named Web log mining. Web usage mining has important effect on improving structure and performance of Web sites, enhancing network security, providing personalized service to users and developing electronic business. It has wide development future.The intelligence computation is an interdisciplinary study involves the physics, mathematics, the physiology, the psychology, the neuroscience, the computer science and the intelligent technology and so on. At present, the intelligence computation technology obtained the widespread application in the interdisciplinary studies of nerve information study, the biological information study, chemistry information study and so on. The progressive of the technology can further promote the development of interdisciplinary studies of nerve information study, biological information study, chemistry information study and so on. In turn, the latter research and develops will also greatly promote intelligence computation technology. So the research of the intelligence computation technology has the vital significance thoroughly.The intelligence computation technology decrypts the object used by the specific mathematical model and makes it become a discipline which may operate, programmable, calculate and the visualization. It has the parallelism, auto-adapted, auto-studied. It can excavate rules and discover knowledge in the magnanimous data of nerve, biology and chemistry. It considers the instantaneity and the agility in entire computation process from beginning to end. It can get the satisfying solution in the limited time through decomposes or the transformation the duty.Intelligent Computation is methodology. It shows the capacity of adaptation and handling the new situations. It enable the system to have the inference attribute of exudes, the discovery, the association and abstract and so on. The output of computation intelligence system usually includes the prediction and/or the decision-making.Now the questions of large-scale optimized question and other complex estimation problem bring the intense challenge to the tradition optimization computational method and artificial intelligence technology. People proposed a kind of new computation intelligence method: Evolution algorithm based on evolution thought. Evolution algorithm does not rely on the domain knowledge, is not limited by searched spatial and has the superiority of intrinsic parallelism. Therefore, evolution algorithm gets more and more many applications in each domain. In the evolution algorithm, the genetic algorithm has the model representation, the use are also most widespread.Genetic Algorithm has the very good performance of searching and easily parallelization. It is fit for saluting the optimized question. So the thesis will use the genetic algorithm to data mining based on web log.In this thesis, the work includes studying the basic theory of Web Log Mining System, analyzing the phylogeny, actuality, content and the problems. The genetic algorithm, one of the intelligence computation technologies, is selected as our data mining technology to resolve some problems.The paper presents Web Log Mining System based lots of data mining technology, focus on the design and implement of the Pre-processing module, the Pattern-mining module and Genetic Algorithm of the Pattern-mining and realizes the incremental mining based on new information of Web log.At last, It performs the test based on scheme proposed in this thesis. The test proves that technology solution used by this thesis can discover the frequent user access patterns effectively, modify the access patterns dynamically and predict the user behavior based on the patterns. It can improve the rate of accuracy and recall, improve on the Web and realize the intelligent Web to provide the personal service.