Clustering Of Web Streaming Data Mining Research And Applications

The streaming data according to became the recent years more important data type gradually, streaming data according to is massive, is fast, when changes, cannot know in advance, streaming data according to the inquiry result is approximate, only can reflect the data the overall characteristic, streaming data of datum inquire treatment technology needs reconsidering , this has opened up a brand-new research direction. In addition, development of Internet, make web website must follow interest , visit frequentness of user, visit time and adjust the page structure dynamically, in order to meet user's demands. It is not merely the course of obtaining information that the data are excavated , pay close attention to finding potential and useful knowledge even more. Data excavate technology apply to web flow analysis of data , raise web flow practicability that data inquire. And how carry on valid cluster among flowing data, it is one that attract researcher very great problem of attention.The data cluster flows to excavate the problem in the main research web of this text. First of all , streaming in the data cluster and excavate algorithms after studying web. The data cluster's algorithm streaming in the present model, their time complexity is low, prescroption is good, have good expansibility, but all algorithms based on space of Europe, can only be suitable for the attribute data of pure number value, can't be suitable for mixing the data of attribute , but the data in web application often mix attribute . This text has proposed one kind is not that cluster's technology of the datum flows in space of Europe: Flow to analyse and process to web method to use fuzzy mathematics go on cluster's analysis data, called the analysis of fuzzy cluster. And it is for web classify measure degrees differents for flowing data,last Europe attributes quantitative from and Kazakhstans graceful the smooth from tolerance; Determine the nature attribute can adopt from hamming to tolerance. Web flows the key step of the fuzzy cluster's algorithm of data is two steps: First, use minimum to carry on cluster from cluster algorithm , form one getting initial. Second, until on the basis of minimum to carry on from cluster algorithm what cluster get initial to form a cluster, use cluster's method of the density to get together or cut apart , make cluster's set steady. Finally, adopt some teleeducation true daily record datum of website test, the cluster excavates user's colonies, carry on pattern analysis and rule and find to kind excavated , offer decision for improving website's structure, for carry on web frequent visit of daily record route excavate , web page cluster excavate initial data of offering. Practice has proved: This algorithm is effective , fast.