Font Size: a A A

The Research And Application On Mining Model Of Fuzzy Association Rules

Posted on:2012-12-17Degree:MasterType:Thesis
Country:ChinaCandidate:Q H DangFull Text:PDF
GTID:2218330338456335Subject:Detection Technology and Automation
Abstract/Summary:PDF Full Text Request
In recent years, data mining has been the most dynamic domain in the research of artificial intelligence and database, and has been caught large attention by the experts and scholars. Mining association rules is an important research topic of data mining, which has got desired results in many application areas. In the production of new dry-process cement, each section makes an impact on the quality of clinker, in this paper, the control of calciner in preheater system is taken as an example, and fuzzy association rules is applied to analyze the history data, by mining out the production rules, the expertise database is enriched and guidance is given to the real production optimization.On the basis of research results at home and abroad, current fuzzy association rules is introduced, according to its shortcomings, a new algorithm is put forward in this paper, which is applied in the control of calciner, and validity and feasibility is analyzed.The main research work of this paper is shown as follows:(1) First, in the process of fuzziness, an improved clustering algorithm called DFCM is brought forward. On the basis of the conventional fuzzy-c-means method, in connection with its faults of been sensitive to the initial cluster centers and easy to fall into local minimum, density function is introduced to get the initial centers, which overcomes the effecting of cluster centers to the rules, and increase reliability of the rules.(2) Next, in the process of mining the rules, based on the Apriori algorithm, according to the shortcomings of increasing consuming time while scanning database repeatedly, a gradually pruning method is put forward, which can reduce the number of scan and raise the efficiency.(3) At last, practical application of this algorithm is introduced, and is applied in the control of calciner. After representing the related concepts, the history data and key factors are used to mine out the rules, also the rules are proved to be feasible and valid.
Keywords/Search Tags:data mining, fuzzy association rules, fuzzy clustering, calciner
PDF Full Text Request
Related items