The Cluster Analysis Technique Applied Research, Customer Consumption Patterns In China Mobile

In today's society with rapid development of science and technology, customer is the security of enterprise's survival and development. If we own customers, we will own the largest market. Maintaining customers, attracting customers and fully exploring clients' earnings potential scientifically is the key of improving the core competitiveness of enterprises. And the cluster analysis for clients is just a preparatory work to get, stabilize and increase the profitable customers.Cluster analysis also known as group analysis, which is a statistical analysis method to study the (samples or indicators) classification. With the increasing demanding to cluster, sometimes just depending on experience and professional knowledge is difficult to classify categories exactly, people gradually put element analysis technology into numerical taxonomy and form cluster analysis. As an important topic of data mining, cluster analysis technology can divide data into natural groups, and describe the feature of each group. It can explore the valuable information and provide a scientific judgemental basis for the understanding of the follow-up data. Cluster analysis is very rich in content, including systematic clustering method, ordered sample clustering, dynamic clustering method, fuzzy clustering, graph theory, clustering method, clustering prediction method, etc. By researching and analysising clustering results, operators can provide a necessary service and take targeted marketing strategies in accordance with behavior characteristics of different customer groups, and discover the potential needs of clients and ultimately bring greater profits for the company.Facing the increasingly fierce competition of communication operators, China Mobile also needs an efficient clustering analysis technique to provide a basis for making customers marketing decisions to enhance the value and satisfaction of customers, increase the efficiency of enterprises. Therefore, to study how to use cluster analysis techniques to achieve the clustering application of China Mobile customer consumption patterns is very important.Based on an analysis of clustering technique, first, we do pre-processing work for China Mobile clients data and get an aggregate data applied to the algorithm model; Then, use one of the traditional algorithms-the K-means algorithm to analyze the mobile customers and achieve effective results; For a number of deficiencies of K-means algorithm, we establish a analysis model of ant colony algorithm and fuzzy clustering algorithm, combining the advantages of the two algotithms, get a valid clustering results. Experimental results show that, on the one hand, this algorithm effectively overcome the sensitivity of initialization of the fuzzy C means algorithm by using ant colony algorithm robustness (stability); on the other hand, it can accelerate the convergence by parallel distributed computing and improve clustering efficiency. Furthermore, clustering results not only contain "either-or" conclusion, also get "and also" features of samples through the fuzzy C means algorithm, reflecting the associated extent of one customer with each category, it provide more basis to analyze the real features of customers to find potential customers and changing customers. Finally, according to the needs of different type customers and China Mobile's business content, we make a series of feasibility packages for customers, it has certain guiding meaning and applicable value for China Mobile companies to draft customer marketing strategy.