عنوان مقاله [English]
In today's competitive world, integrated systems constitute an essential need in organizations. Managers providing such services are mainly concerned about identifying loyal customers and establishing long-term relations and a sustainable profitability. This study aimed to assign support forces to answer customer calls and requests in order to gain their utmost satisfaction. First, research data was collected from customers of Chargoon Co. in Iran, and the customer behavior was identified and analyzed. Then, using decision tree algorithms such as: random forest, gradient boosted trees, random tree, CHAID, C5, and ID3, a number of trees were extracted based on accuracy, classification error, weighted mean recall, and weighted mean precision criteria using Rapid Miner features. Next, the best trees were identified based on mean of quality assessment and accuracy of over 67%, classification error of under 27%, weighted mean recall of over 62%, and weighted mean precision of over 66%. Then, rules from these trees were extracted and analyzed. Finally, the extracted rules were examined and ranked based on expert views. The results indicated that the extracted rules are reliable and can be implemented in the organization.