Research On Bank Customer Clustering Based On Improved K-means Algorithm

With the rapid growth of C hina’s economy in recent years, competition in all walks is getting increasingly fierce.Especially for banks and other financial fields, how to survive in the intense competition environment has become the most concerned problem. In recent years, the development of information technology has brought great changes to the competitive environme nt of enterprises. Enterprises have gradually changed their central focus to the customer, though products and services were their all focus in the past. Enterprises gradually realize that if they can effectively master the customers, they can effectively grasp the good performance. If they can meet the needs of customers more timely, they can be more able to meet the needs of the market and to stand out in the industrial competition. At present, most of the enterprises in our country have established the internal customer management system. O ur financial industry has accumulated a huge amount of customer data resources. If they can effectively understand and use these customer data, it will be of great benefit to improve the service level of enterprises. A mong them, the accurate classification of customers can make the enterprise more effectively in providing more targeted and more effective services to different customer groups. Banking industry already has a collection of massive data sets today. But how to effectively use the existing information and digging out the real decision- making value information for bank decision makers are important research topics.It is clear that with the exponential growth of data, the traditional way of artificial custome r classification is unrealistic. As to the customer classification accuracy, traditional domain expert manual classification is better than automatic classification; But as to the classification efficiency, the efficiency of automatic classification is muc h better than that of artificial classification. The cost of continuing using artificial classification to classify bank customers is high and unrealistic.Data mining technology can discover hidden, effective, valuable and understandable patterns from a large number of disordered data, and then find useful knowledge, obtains the time tendency and the correlation, and provide decision makers with the ability of decision support. It provides a new significance for the processing of massive data accumulating in the information age. Along with the development of data mining technology, as an important part of data mining technology, clustering technology has been widely used in many aspects, such as data analysis, text analysis, image processing and market pred iction. C lustering algorithm has become a very active research topic in the research of data mining technology. After many years of research, many clustering algorithms have been proposed. K-means algorithm is one of the most widely used clustering algorit hm based on square error iteration. However, the original K-means clustering algorithm have many defects.This paper focuses on the improvement of the traditional K- means algorithm and its application and implementation in the direction of bank customer classification. Firstly, the basic theory of data mining is introduced. Then it introduces the basic theory of clustering analysis algorithm, introduces the K-means clustering algorithm in detail, analyzes the advantages and disadvantages of the algorithm and puts forward the improvement of the algorithm. After that, we define the theoretical system of bank customer classification to provide theoretical support for the bank customer clustering. Finally, based on established the bank customer classification model, this paper conduct the actual bank customer clustering experiments on the detailed transaction data from the financial department of a bank in 2012-2015.Finally, the clustering results are analyzed and summarized in detail.