Font Size: a A A

Research Of Algorithms On Association And Clustering In Business Database

Posted on:2006-09-13Degree:DoctorType:Dissertation
Country:ChinaCandidate:Z WangFull Text:PDF
GTID:1118360155453608Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the development of computer science, computers and related technology has been the basis of the management of enterprise, and the abilities of generation, collection, storage and processing improve greatly. Because of the surprisingly increment of data, the traditional methods fail into the difficulties, especially in the domain of business, telecom, internet, and science research. Data Mining is brought forward for surmounting the above difficulties. It is a newly-generated subject which utilizes the knowledge of machine learning, pattern recognition, database, statistics, and AI. Business intelligence (BI) converts all kinds of information into what enterprise interests, express it for helping make decision of enterprise, and strengthen the advantage of competition in market. Dataware, OLAP, and data mining is the basis of BI, and mining association rules and clustering is the most important fields of data mining. A data stream is a continuous, huge, fast changing, rapid, infinite sequence of data elements. Analysis model for static data for e-business has developed a lot,however, analyzing stream data in e-business just begins developing. In this paper, we introduce the implement of mining association rules and clustering in data stream. Through analyzing the situation of streaming data development, we focus on the research of mining in data streams. This paper also discuss the below aspects: (1)We introduce the related work of e-business, mainly present the related knowledge of data mining, and systematically summarize the famous algorithms concerning mining association rules and clustering in data streams. According to the sparse characteristic of data of supermarket, we implement algorithms for mining association rules from transactions, and optimize them by summing and leveling the data of supermarket. We also implement the clustering algorithms which debase the computer complexity and exclude the...
Keywords/Search Tags:Business intelligence, Data mining, Data stream, Association rules, Clusterig, MFISF, Two-tier, Regressive strategy
PDF Full Text Request
Related items