Comparing data mining and fuzzy logic techniques to identify behavior of customers

This study compares data mining techniques and fuzzy logic for Identify customer behavior and voice of the customer to be used in the process of target costing.
In this research, the data relating to sales in the data warehouse of Farzad tile producing company in years 2014 and 2015 have been used. The results of the test of hypotheses suggest that the rate of correct prediction for customer features in neural-fuzzy networks with activation function, Fuzzy Clustering is 0.941and in Multi-layer Neural Network with sigmoid activation function is 0.927 and in the multi-layer neural network with tangent function is 0.882 and in Radial basis function network with Max softball function is 0.918. The results show that fuzzy neural network has better results than other methods used to predict the characteristics of the target customers.