Font Size: a A A

Multidimensional Association Rules Mining

Posted on:2002-04-08Degree:MasterType:Thesis
Country:ChinaCandidate:G J SongFull Text:PDF
GTID:2208360032456826Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Data mining is the process of discovering interesting knowledge from large amounts of data stored either in database, data warehouse, or information repositories. The mining of the association rules is one of the most important issues in the field of data mining. In classical association rule mining, records in a transaction database contain only items and are identified by their TIDs. Althougb transactions occurred under certain contexts, such as place. time, customer, etc., such contexts have been ignored in classical association rule mining. We call this kind of rule as multi-dimensional association rule. It can provide useful information about our world. and the study about it will be much meaningful work and also have broad application. In this paper, we discuss two kinds of question about multi-dimensional association rule in i ni ng: intra-transaction association and inter-transaction association. The following will be given the scheme. To the first question~ we propose a novel MFP-tree (Multi-dimensional Frequent Pattern tree)structure, which consists of DP-tree and FP-tree for storing information about dimensional patterns and frequent item sets respectively, and develop an efficient MFP-tree-based algorithm for mining Multi-dimensional Association Rules (MAR). The algorithm we presented here follows the paradigm of divide-and-conquer. Our performance study shows that it is efficient and scalable for mining MARs. To the second , we introduce the notion of multi-dimensional inter-transaction association rules, study their measurements: support and confidence, and develop algorithms for mining inter-transaction associations by extension of Apriori. For ease of understanding and simplicity, the discussion has been limited to one dimension. Further extensions of multi-dimensional inter-transaction association rules and potential application are also discussed.
Keywords/Search Tags:Data mining, Multi-dimensional Association Rules (MARs), inter-transaction association, intra-transaction association
PDF Full Text Request
Related items