Research Of Clustering Algorithm And Its Application In Mobile Customer Churn Management

As an important area of data mining, clustering analysis is a necessary method of data partitioning. Clustering plays a considerable role in many discipline's areas, such as finance, marketing, telecommunications, industry, biology and medical field. Currently, clustering algorithm is roughly divided into partition-based approach, hierarchical-based approach, density-based approach, grid-based approach, model-based approach and fuzzy clustering.With the rising profits of mobile enterprises, customer churn is also increasing. The development of the mobile enterprise is seriously affected by the average life cycle of customers. Therefore, the mobile communications companies urgently need an effective method to determine the customers who are possibly lost, then make the appropriate retention policy to reduce the enterprise's economic losses.The paper studies the clustering algorithm carefully and puts forward a new method formed by density-based approach, partition-based approach and hierarchical-based approach, which is applied into the segmentation of mobile customers. What's more, find out the customers who are possibly lost and their consumption characteristics by means of the clustering techniques. Last but not least, some corresponding marketing advices should be proposed for these customers.