Clustering analysis with particle swarm optimization and data mining methods: Application of a cluster of customers of a company established in Izmir

Nowadays, for the institutionalization of companies, it is necessary to provide specialization in their fields by giving them a good understanding of the customers in a number of methods, acting according to customer characteristics and giving their employees the tasks in this direction. However, this is not easy with traditional methods in companies with a large number of customers. In this thesis, the customers were clustered with both K-means and Two-step Algorithm Methods from Data Mining Methods as well as the modern heuristic technique, Particle Swarm Optimization Method. The concepts related to these methods have been explained and a comprehensive literature search has been done. SPSS Clementine 10. 1 has been preferred for K-means and Two-Stage Algorithm Methods. The Particle Swarm Optimization Algorithm has been used for this purpose in the MATLAB program, adapted to the cluster. Dunn Index is used as the objective function in the algorithm and cluster results are obtained. Keywords: Data Mining, Clustering Methods, Clustering with Particle Swarm Optimization, Customer Clustering.