International Journal of Computer Science Issues

More than a traditional journal...

In todays competitive environment, operators are investing in understanding their customers better, especially their most profitable customer groups and the groups that have the biggest potential to become such. By segmenting customers based on their behavior, operators can better target their actions, such as launching tailored products and target one-to-one marketing, to meet the customer expectations. Currently, telecommunication operators build segmentation models based on information combined from different sources such as billing data, call detail records (CDR) and customer surveys. The problem is that the gathered information is of multidimensionality, which results to poor accuracy when running a model on whole data set. Therefore, there is a clear demand for Business Intelligence (BI) to help in gathering and reduction of dimensionality to analyze and visualize data and information concerning business entities to help them make a better-informed decision concerning their business. With the help of data mining techniques, segmentation can be done automatically and based on actual customer behavior. This study investigated customer behavior segmentation among mobile service providers using K-means algorithms. The general objective of the study is to provide customer behavior segmentation in mobile telecommunication markets using K-means Algorithm. The specific objectives include to handle multidimensionality data using K-means algorithm with Principal component analysis, to determine the value of parameter K (number of clusters) using stability plot before clustering, to use finanacial variables (mean monthly charges) for each frequently used service as inputs in k-means for Segmentation, to evaluate Clustering results and determine the most profitable segment using completely randomized design (CRD). The experiment to achieve the objectives was being done on R software. Results show that Cluster 3 and 1 are the most profitable segments. The operators often need to design distinguishable marketing strategy based on different behavior of their mobile subscribers in order to improve their marketing result and revenue. A customer life cycle model is suggested considering the past contribution, potential value, and churn probability at the same time.