Exclusive clustering technique for customer segmentation in national telecommunications companies

This study aims to empirically examine consumer behavior based on customer transaction history. Analyzing consumer behavior can provide very useful information for businesses in making decisions, particularly business decisions toward customers, in order to survive in such intense competition. Companies are becoming faster and more precise in reading environmental conditions and predicting what conditions may occur as a result of machine learning technology. This technology can also assist companies in making decisions that are more targeted according to actual secondary data provided for research. One of the machine learning methods, unsupervised learning, can help explicitly identify hidden structures or patterns in data and determine correlations. This method uses the Exclusive Clustering method, using two algorithms, namely, K-Means and K-Medoids, to use the comparison method to get optimal segmentation results. The results obtained are expected to be a reference for making a change in the company's marketing policy in order to retain and gain customers who are constantly decreasing.