Font Size: a A A

Research And Application Of Two Improved Association Rules Mining Algorithm

Posted on:2018-05-15Degree:MasterType:Thesis
Country:ChinaCandidate:P LinFull Text:PDF
GTID:2348330515989578Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
With the rapid development and wide application of database technology,large amounts of data have been accumulated.How to use those data effectively and reasonably,and how to dig into the knowledge hidden behind the data,have become the focus of attention.Traditional statistic data analysis no longer satisfies current needs.Instead,data mining,as a new tool to analyze data,is receiving more and more attention.It is now a hot research field in information technology.As an important branch of data mining,association rule has taken quite great development and made important achievements in many aspects.However,association rule still faces many challenges when there are massive amounts of data.On its path of development,researchers put forward various data mining algorithms,including Apriori algorithm,FP-Growth algorithm,and DHP algorithm.While integrated with computer technology and practical application scenarios,many algorithms have plenty of room for improvement.This paper concludes two improved association rule mining algorithms based on the analysis of current association rule studies both in China and abroad.This paper mainly focuses on the following aspects.(1)We classify and summarize the basic theories about data mining and association rules,and introduce two classical association rule mining algorithms,namely Apriori algorithm,and DHP algorithm.(2)We come up with an improved algorithm based on Apriori algorithm and binary system.This algorithm transforms the transaction database to decimal numbers,and then use AND operation to count the support of candidate sets.Finally,experiment data verifies that the improved algorithm requires less runtime.(3)Based on DHP algorithm,we use red-black-tree data structure to deal with the conflicts occur in the process of hashing,which allows every candidate set to be count individually so that DHP algorithm no longer need to scan the database repeatedly to get the support of each candidate set.(4)Use the DecBitApriori algorithm to tap the supermarket shopping data,put forward the relevant advice on the placement of supermarket goods.
Keywords/Search Tags:data mining, association rule, Apriori algorithm, FP-Growth algorithm, DHP algorithm, Commodity display
PDF Full Text Request
Related items