Font Size: a A A

Research Of Data Extraction Based On Quotient Space Granular Computing For The Click-stream Data Warehouse And Mining Algorithms

Posted on:2012-06-19Degree:MasterType:Thesis
Country:ChinaCandidate:Y J WuFull Text:PDF
GTID:2178330335466791Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the rapid development of e-commerce sites, a large amount of click stream data in web site come into existence every day. The click-stream data contains a lot of very useful information, such as the source of customers, consumer behavior, the access interest and so on. Analysising these data effectively, not only do play a guiding role in enhancing the degree of adhesion sites to the construction of e-commerce sites, but also to optimize the site topology further and increase product sales, improving the site quality of service, creating value for enterprises .This article builds upon a process which is based on a shopping site of the clickstream data warehouse. Firstly, outlining the basic concepts and related technologies of the data warehouse and data mining, and then with web knowledge of click-stream data warehouse, from the overall data warehouse architecture, model design, ETL and other aspects of shopping sites to build the clickstream data warehouse described the detail of the build process, reaching a clickstream data warehouse available solutions, and building data warehouses using the Microsoft tools were clickstream data warehouse implementation. Followed using granular computing method of collection of clickstream data to discuss the theory of from the multi-level point of view of mining data sources, streamlining data collection processes, analyzing log files from the web server access problems and clickstream data preprocessing and some of the difficulties presented quotient space granular computing theory the idea to collect clickstream data, the method can access to click-stream information efficiently and flexibly , avoiding the redundancy of the data preprocessing. Then using the association rules of the web data mining, a user preference path algorithm of mining frequent is proposed, using the algorithm can find the user's interest in the common view, understanding the user's behavior from the multi-depth side, and then guiding the sales site to improve its topology. Finally, online analytical processing the process output data to click-stream data to demonstrate the value of clickstream.The construction project of the clickstream data warehouse can not only provide website information analysis, but also support the deep-level digging and sales analysis on the basis of the log data warehouse usage patterns.
Keywords/Search Tags:Data Warehouse, ETL, Quotient Space Granular Computing Model, Hierarchical User Structure Model, Web Log Mining, Browse interestingness, Granular Computing Association Rules
PDF Full Text Request
Related items