The Study of Applying Market Basket Analysis to the Most Appropriate Product Mix on E-Commerce Marketplaces

摘要

CRM is the process of controlling efficient, cost-effective flow knowledge to individual customers' requirements, and interaction with customers through channels for the purpose of escalating the customer's lifetime value. In other words, product is the core business process of traditional marketing, while the customer is the core business process of CRM. The former concerns which product is the best sales while the latter does which customer is faithful. In order to digest the individual customer, We should set up customer's perfect database which contains data of interaction with each customer, such as primary information, transaction, service, activities and the like. This study uses market basket analysis of data mining technique, trying to establish synthetic rules from customers' transaction data. From generating the co-occurrence matrix for single items, generating the co-occurrence matrix for two items and three items, in order to design competitive product mix that convey more convenient shopping value for customers. It is not common to use market basket analysis in real transaction data in domestic study. The purpose of this research attempts to apply Market Basket Analysis models to describe the effectiveness of the criteria to verify its result.