An Application of Artificial Neural Network for Constructing Service Quality Evaluation Model--Telecommunication Industry

• Main Authors: Yung-Hung Wu, 吳永宏
• Other Authors: Chuang Tu
• Format: Others
• Language: zh-TW
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/82105190374263143853

碩士 === 國立臺北科技大學 === 生產系統工程與管理研究所 === 89 === According to the statistics of the Directorate General of Telecommunications Ministry, the number of cellur phones is increasing rapidly. It is fourteen million and six hundred and eighty thousand from that the government opened up the civil proprietor the cellur phone operating to June, 2000. Due to telecommunications’ keen competition, the number of proprietors has decreased from seven to six. The of telecommunications proprietors seldom pay attention to the service quality but to the elasticity rate, the premium plan of the cellphone, additional value, the free charge for setting up, and guarantee money, etc. However, Weng (1996) pointed out that the key factor for the successful corporate is service quality. Thus, this study is not only trying to discuss service quality from the point of consumers’ demand but also consider the recognition of service quality from the point of telecommunication corporate. However, some issues about the development of service quality evaluation model still should be concerned. First, it is important to determine whether the dimensions of traditional service industry are suitable for information industry or not. Besides, we adopt the statistical method to analyze data preciously, but it may not be the best method. According to Schaffer and Green(1988) research, they pointed out that it may reverse the original pattern from adopting factor analysis, and it is not necessary that the results from factor anaylsis on the original pattern are gained. Thus, the statistical method and the neural network method (Back-propagation Network; BPN, General Regression Neural Network; GRNN, Radial Basis Function Network; RBFN) on original demesions without factor analysis and on service quality dimensions with factor analysis are employed in this study. The computational results showed that GRNN with the original demesions has the better learning effects, and its MSE is the smallest. Therefore, this research employed it as the framwork for service quality evaluation. And this is the very first time to integrate service quality evaluation with Internet. The business and consumers can communicate with each other without the time and space limitations through user-friendly interface. Keywords : Service Quality, Statistical Method, Neural Network, Internet