Maziyar Grami - Department of Computer, Technical and Engineering College, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran
Reza Gheibi - Department of Computer, Technical and Engineering College, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran
Fakhereh Rahimi - Department of Computer, Technical and Engineering College, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran

Today, development of internet causes a fast growth of internet shops and retailers and makes them as a main marketing channel. This kind of marketing generates a numerous transaction and data which are potentially valuable. Using data mining is an alternative to discover frequent patterns and association rules from datasets. In this paper, we use data mining techniques for discovering frequent customers’ buying patterns from a Customer Relationship Management database. There are lots of algorithms for this purpose, such as Apriori and FP-Growth. However, they may not have efficient performance when the data is big, therefore various meta-heuristic methods can be an alternative. In this paper we first excerpt loyal customers by using RFM criterion to face more reliable answers and create relevant dataset. Then association rules are discovered using proposed genetic algorithm. The results showed that our proposed approach is more efficient and have some distinction in compare with other methods mentioned in this research.

Feature selection; Genetic algorithm; Data mining; Association rules, customer relationship management