Research For Method Of Market Segmentation Based On Self-organizing Cluster

Market Segmentation is an important step in STP and Customer Relationship Management. STP and CRM are base on a good and effective market segmenting result So, a good segmenting is very importantTraditional Cluster Analysis has two disadvantages applying for Market Segmentation: 1. Cluster Analysis can not automatically get the optimal cluster; 2.Traditional Cluster Analysis could not get a good clustering result when the data and dimension are big.In this paper, we not only researched the Self-organizing Cluster deeply, but also improved the deficiencies in the old arithmetic. At last, I applied the improved arithmetic to the market segmentation of Shifang tobacco factory. Comparing with statistic Cluster Analysis, the Self-organizing Cluster can automatically get the optimal cluster without people's attending. It can reduce subjectivity in the clustering process. The major achievements in this paper are as follows:1. Applying Self-organizing Cluster to Market Segmentation, including designing operation steps, programming and demonstration. the Self-organizing Cluster can automatically get the optimal cluster, fit in with big data space.2. Developed the Self-organizing Cluster, solved some questions of the old arithmetic. The new Self-organizing Cluster can use in market segmentation better. The achievements are as follows:? Brought forward the new consistency criterion. Comparing with the old consistency criterion, the new one can explain why smaller criterion, better clustering.(2) Brought forward the method of splitting dipole. Using it, we can get the optimal clustering result when the minimum consistency criterion is zero, the old arithmetic is invalidation.(3) Inducted Gower Distance. It makes the arithmetic fit market segmentation better.? Programed using VB.3. Researched the difference between Self-organizing Cluster and Cluster Analysis and the difference between new Self-organizing Cluster and old Self-organizing Cluster.4. Used the data of Shifang tobacco factory marketing to demonstrate.The research of this paper is the development of market segmentation method. It has practice significance to how to get an optimal consistency clustering.