Applying Artificial Neural Network to Customer Loyalty Analysis －Case study on Health Examination Business

• Main Authors: Ching-Tan Yang, 楊清潭
• Other Authors: Ting-Li Lin
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/38eeub

碩士 === 銘傳大學 === 管理學院高階經理碩士學程 === 94 === In recent years, customers value has become the focus of business administration. The issue of Customer Relationship Management has been gaining a lot of discussion and research attention. Enterprises are investing many resources to enhance understanding of customers and establish good relationship with customers, in a hope to raise customer satisfaction and customer royalty and eventually to obtain more groups of royal customers. According to the 80/20 rule, we assumed an enterprise should use major resources to keep 20% of the customers who have best fidelity, and put minor resources on the rest of customers. Differentiating customers by the royalty level and providing diverse services along with the royalty degree, help to attract more royal clients and bearing high willingness to recommend new consumers subsequently, so as to gain optimal profit for the customer service oriented enterprise. In this research, we investigated past transaction data of customer’s health exam to perform customer royalty analysis. We applied Jones & Sasser’s (1995) basic behavior, re-purchasing enthusiasm and derivative manners as the basis for the royalty measurement. Additionally, attributes related to the medical industry, such as health status, consuming period and population variation were also used as input parameters for the artificial neural network. Based on the RFM analysis model proposed by Hughes (1996), we took individual distinction as weighted factors to rectify the problem that industrial diversity could cause different significance of RFM attributes. The RFM attributes were then transformed into a three-dimensional vector to compute the absolute distance as the customer royalty index. Next, with the neural network technique, a data mining technique with learning ability, we have past customer transaction data as input and current customer royalty indices as learning goal to produce a neural network model to predict future customer royalty indices. An effective screening model for customers with high re-consuming potential is established.