Clustering Customers of bank Refah branches for combining genetic algorithms and C- Means in a fuzzy environment

Today, one of the major challenges that planners and managers are grappling with in the field of banking services is customer recognition, and distinction between different groups of customers. It is obvious that the use of appropriate model gives the bank the opportunity that fits your valuable suggestions along with demands for targeted sectors and provides design and thus improves bank performance from different perspectives. The aim of this study is using and appropriate model for clustering customers based on indexes including novelty, number of transaction and financial factors. In this paper, for clustering data, the genetic algorithm combining with fuzzy C-means is used to overcome problems such as being sensitive to the initial value and getting trapped in the local optimum. The simple random sampling method is used to obtain the sample. The findings show that the first cluster of customers due to its high performance in "novelty", "number of transaction" and "financial factors" index are loyal customers and the second cluster of customers because of low performance in "novelty" index, mean performance in "number of transaction" index and high performance in "financial factors" are among those customers who are turning away from bank.