Research On Electric Water Heater After-Sale Service With Case-Based Reasoning And Customer Knowledge Management

Because of the home appliance industry product homogeneity and the characteristics of a large market share, which attracting a large number of capital to enter and lead to intense competition within the industry. As technology advances, competition among home appliance industry, the product competition gradually changed from the original service to the after-sale service competition. After-sales service has become an important part of the brand, also an important consideration for most consumers to buy household electrical appliances. Now, solving after-sale service by people is the most common method to appliance industry, such as setting up service centers to manual mode to accept phone about repairing, complaining and consulting. In this mode, it is high repetitive, time-consuming and labor-intensive to solve after-sale service problem, meanwhile can't ensure customer satisfaction. In this situation, it is urgent for after-sale service of the home appliance industry to have the case-based management application, and the method of case-based reasoning (CBR) just to meet current needs. On this basis, we can apply the concept of customer knowledge management in it, which can improve and enhance existing service levels.The essay selects the most typical household appliances——electric water heater as the research objects, and uses the method of CRB to construct the after-sale system of electric water heater management system. The major steps are the analysis of the knowledge composition, case representation, case retrieval, case correction, learning and case base maintenance. Firstly, through research and interviews, the content of the water heater service carried out a detailed analysis, summarized the main contents of service types, and having a foundation of the case representation. About case representation, because there is a structural framework of representation, inheritance, and natural of advantages, especially for complex case representation and the representation can express not only qualitative attributes, but also quantitative properties. So the essay chooses frame method as case representation's method. In order to improve the efficiency and accuracy of search, the order of case retrieval is index, keyword and K nearest neighbor. The advantage of three methods 'combination can deal with indexing attributes, text attributes and quantitative attributes. Gradually reduce the case library to retrieve the search range, thereby enhancing the efficiency of retrieval. The attribute weights in K nearest neighbor are obtained by experts and AHP. Case correction mainly uses direct correction and the secondary correction. The direct correction method can't meet the need of new after-sale service of new products, so it can use secondary correction to solve these types of problems in order to improve the accuracy and reliability of correction. In the Case base maintenance process, proposing the four types of problems of it. It is inconsistent cases, redundant cases, incomplete cases and garbage cases. In solving these issues, through the utility's calculations, the similarity comparison and important feature of the calculation of missing values, and finally by removing the way to solve these four issues of case library, to improve the system performance objectives.In the process of the electric water heater with CBR, the essay adds the concept of customer knowledge management. For example, when expressed in cases such as the construction of two sub-frames based on the client, in the case retrieval and customers can also be related to the keyword search. In the KNN method when determining the weights of attributes and customer-related weight is relatively high, embodies the concept of customer satisfaction oriented.In the part of system implementation, this case knowledge of storage and preliminary implementation of the system use SQL Server 2000 and VB 6.0. The important part is case representation and case retrieval module. In case representation the data table is used, case retrieval through the use of index, keyword matching, SQL query and calculation of K nearest neighbor method to execute the searching function.Finally, testing of the system verify the search algorithm.