Customer Segmentation Using Particle Swarm Optimization and K-Means Algorithm

Customer segmentation is an implementation of clustering in the data mining process. Customer segmentation divides customers into certain classes to help a company to understand each customer. This paper analyzes 33,441 rows of a transaction data and transforms it into 914 rows of Recency, Frequency, and Monetary data (RFM) to identify potential customer. Clustering method uses are the combination of Particle Swarm Optimization (PSO) and K-Means algorithm. The combination of these algorithms aims to take advantages of both algorithms and remove their weakness. K-means is very sensitive to initialize the cluster center because it does randomly. PSO is uses to optimize the cluster center and help K-Means to cluster better. The clustering experiment uses several numbers of a cluster. The best numbering of the cluster for this experiments are two clusters according to Davies-bouldin Index (DBI) method.