服務系統之發展已成為未來趨勢,在服務資源有限且不易保存的特性下,如何有效的利用服務資源,減少浪費,是社會及企業所關心的問題,而企業常藉由預約系統,來做為提昇服務資源利用率的工具, 而決定預約系統優劣的重要關鍵,既在於顧客履約預測的準確性。 本文在研究手法上,是以顧客面為導向,探討分析顧客失約的行為特徵模式,找出要因參數,建立顧客屬性分類及考慮過預約 (overbooking)策略之訂定,並以容錯力較佳之類神經網路做為預測運算之工具,在本文案例中,以本研究所構建之顧客履約預測模式,經測試(MSE=3.033),較傳統手法預測(MSE=6)結果為佳。 在服務業中,服務者的服務成本及被服務者的等候成本,都是很值得研究及關心的問題,企業為求永續經營,創造利潤,選擇最適當的服務規模,來提昇企業之競爭力及服務品質,才可創造雙贏的局面。
The development of service system has already become the tendency of the future. For the reason of service resource constrained and not easy to keep it, so how effective to use the service resource and reduce the expenses , is the most concern for our society and the enterprises, by means of appointment system, the enterprises promote service resource utilization, and how to decide the crucial point of the appointment system’s quality is up to the accurate of customer’s fulfill appointment prediction. The way of researching about the prediction of customer’s fulfill appointment in this thesis, bases on the customer’s behavior, then I discuss and analyze the customer’s behaving model of breaking on appointment, to find out the significant parameter, set up customer’s classification and consider the rules of overbooking strategy, then using neural network, which has the high allowance of errors to be the tool’s of prediction. After analyzing in the real case, we found that the predictive model of customer’s fulfill appointment of this research (MSE=3.033), is better than tradition’s analyze prediction (MSE=6). In service industry, the service cost and waiting cost for both customer and enterprise are worth to be researched and concerned. For running business in a long time with high profit, the enterprises have to choose the most suitable service’s scale and service’s quality, in order to promote business’s competition, and then reach the situation of two- wins.