Revealing a Novel Method for Detecting Positive and Negative Optimal Performance Association Rules in Very Large Databases Using BPSO
Association rules mining is one of the useful data mining algorithm in presenting meaningful information through database. One of the important challenges for association rules mining is that; it might be extracted millions of rules which mostly are idol, furthermore current methods only seek positive rules which finding out the negative ones is more prominent. In this method we mingle data mining and Evolutionary algorithms including association rules, Particle Swarm Optimization, whose goal is discovering the pattern and positive, and negative rules and also optimum one through large database, also this algorithm is capable to present scarce rules, which might be neglected by administrator. Consequences by recent algorithms could help administrator in making many resolutions.
Also this algorithms result has been compared with Apriori. The results indicate the algorithms efficiency. Collecting and preparing data in this survey has been performed in SQl server 2005 and algorithm performed in MATLAB software.
Keywords: Data mining, Association rules mining, Evolutionary algorithms, large databases, Particle Swarm Optimization
Download Full-Text
ABOUT THE AUTHORS
Salah Karimi Haji Pamagh
Department of Computer Engineering, Science and Reseach Branch Kurdistan, Islamic Azad University, Sanandaj, Iran
Mehdi Afzali
Department of Computer Engineering, Zanjan Branch, Islamic Azad University, Zanjan, Iran
Amir Sheilkh Ahmadi
Department of Computer Engineering, Sanandaj Branch, Islamic Azad University, Sanandaj, Iran
Salah Karimi Haji Pamagh
Department of Computer Engineering, Science and Reseach Branch Kurdistan, Islamic Azad University, Sanandaj, Iran
Mehdi Afzali
Department of Computer Engineering, Zanjan Branch, Islamic Azad University, Zanjan, Iran
Amir Sheilkh Ahmadi
Department of Computer Engineering, Sanandaj Branch, Islamic Azad University, Sanandaj, Iran