Application Of Cluster Analysis In Port Of Customer Segmentation

With the increasing competition in domestic and foreign ports and port their business development, domestic port enterprises operating mode must be transformed into an information-based, data-centric to customer-centric international advanced model, and to achieve this scientific premise business model is the need to study customer segmentation work. Customer Segmentation Method of Chinese port enterprise is based on statistical or based on experience classification methods, and no real information exchange between businesses and their customers, unable to meet the different services provided strategies for different customer needs.As an important method of data mining techniques, cluster analysis has become a very important in this field of research. Cluster analysis is in the absence of any prior knowledge of the case will be a number of sample data (or variables) according to their degree of closeness in the nature of automatic classification, and ultimately achieve blind classification sample space. Cluster analysis using data mining for customer segmentation, not only can handle dozens, or even hundreds of variables, and thus a more accurate description of the customer and objectively reflect the characteristics of the client group and within a comprehensive reflection of the characteristics of various customers; but also can help marketers have more in-depth and detailed understanding of customer characteristics, easy to implement changes in the dynamics of customer behavior tracking; thus achieving to provide differentiated services to customers, improve customer satisfaction and loyalty, and ultimately create more value for the enterprise.This paper describes the background of the existing port information to explain the necessity of information technology to advance to today’s stage production data of port and the urgent need for the use of data analysis and mining capabilities of mining technology. Then a detailed overview of the basic theory of customer segmentation, cluster analysis method is applied to the basic theory of customer segmentation and cluster analysis algorithm is designed to lay the theoretical foundation of cluster analysis method in Customer Segmentation later article. Analyze the situation port customer database, selection and construction of the port attribute of customer segmentation, and preprocessing customer segmentation necessary data in the port enterprise is to start doing customer segmentation data preparation. Secondly, mainly analyzes the classic K-means clustering algorithm, AP algorithm and Particle Swarm Optimization algorithm deficiency in the port of customer segmentation, proposed algorithm combines the advantages of three kinds of hybrid clustering algorithm. The algorithm uses the AP algorithm to select the value of K, takes advantage of Particle Swarm Optimization algorithm for global search ability and strong K-means algorithm local search ability and other characteristics. Through experimental verification of the proposed method improves the effectiveness and accuracy of clustering algorithm to speed up the convergence rate. Finally, the improved K-means clustering algorithm is applied to management practices in Guangzhou Port Group Co., Ltd. Business. Interpret the results for customer segmentation and analysis features of each type of market segmentation. Combined with the actual situation ports for existing customers, given the appropriate customer marketing objectives and strategies, and proposed the development of new customer markets.