全球汽車數量逐年快速增加,各大汽車品牌為增加銷量,除研發提升車輛安全性及各功能外,也相當重視售維修品質,維修品質可影響其銷售量,所以各車廠對維修品質相當重視。汽車是一種結構零組件相當複雜的交通工具,維修人員除了原廠的通則性技術指導外,對於其他非通則性技術問題,需靠經驗累積或較資深的維修技師才能解決問題,除了維修人員的維修能力不一外,維修知識與經驗無法適時保存與利用,事實上目前車輛進廠時許多車主會針對進廠的原因症狀進行敘述,此敘述也會被記錄下來,因所記錄資料為非結構化資訊,故本研究將透過車廠的維修記錄與車主的問題描述,對於維修準則判斷不易及非通則性維修知識的獲得或是傳承,透過結構化及系統化的方式將其儲存,透過文字探勘技術,進行關鍵因子分析,應用於改善維修決策的建議,配合專家建議及驗證,建立問診知識庫以幫助服務廠維修人員獲得維修決策分析,以提升車廠售後服務品質。
Traditionally, a mechanic often rely on personal experience to inspect a vehicle even though there are some unexpected problems that are out of the original technical guidance. In order to enhance the service quality of vehicle maintenance, this thesis design a knowledge-based mechanism by capturing the problem descriptions of vehicular owners and the diagnosis results of vehicle mechanics to retrieve maintenance knowledge and experience. Therefore, this study uses a text mining technique and data analysis technologies to statistically analyze key factors and suggestions. The experimental results show that these text-mining techniques can improve the maintenance quality and therefore optimize the decision support for vehicle maintenance to enhance service quality.