美髮沙龍顧客消費行為分析-以某家髮廊為例

The Analysis of Customer Consuming Behavior in Hair Salon-A Case Study of a Hair Salon

姚汝臻(Ru-Jhen Yao)

指導教授 ： 陳世穎

共同指導教授 ： 林益永(Yih-Yeong Lin)

國立臺中科技大學/資訊與流通學院/資訊工程系碩士班/碩士(2017年)

摘要

此研究利用支撐向量機（SVM）來進行美髮沙龍店內顧客來店消費資料之預測分析。透過已知的顧客以往消費紀錄等相關資料，包括來店時間、消費項目：洗髮、剪髮、燙髮、染髮、護髮等項目進行預測分析。此研究收集台中某家髮廊的某位設計的顧客資料為實驗數據，將利用支撐向量機（SVM）發展預測模型。研究結果顯示支撐向量機（SVM）可以準確地預測髮廊在顧客來店消費時間上預測分析，分析結果具高度的準確性。實驗結果能幫助設計師對於顧客的回流率有準確的預測分析，且能提供設計師對每位顧客來店之頻率進行統計分析，讓設計師能做有效的應對方案，提高業績的成長。

關鍵字

顧客關係 ； 美髮沙龍店 ； 支撐向量機

並列摘要

This research uses the Support Vector Machine (SVM) to predict and analyses beauty salon customer visits and expenses. Prediction and analysis are made from the known customer’s former record and correlation data, including time of visit, purchased services: hair wash, cuts, perms, dye, care and so on. This research collects data from one design customer in a beauty salon located in Taichung, and uses the Support Vector Machine (SVM) to develop a prediction model. The findings showed that the Support Vector Machine (SVM) may forecast accurately the beauty salon customer’s time of visit, the analyzed result had a high accuracy. The experimental result can help the designer to have an accurate forecast analysis regarding customer's reflux ratio, and also can provide analysis for the designer’s frequency of customer’s visit, and carry on the statistical analysis, enabling the designer to make and execute a plan effectively, and improve performance.

並列關鍵字

customer relations ； beauty salon ； Support Vector Machine

參考文獻

（七） 莊玉玲（2005）。顧客關係管理對顧客滿意度與忠誠度影響之研究－以台灣砷化鎵半導體磊晶廠為例。未出版碩士論文，中原大學企業管理學系，桃園市。

（十二） Cherkassky, V. & Ma, Y. (2003). Practical selection of SVM parameters and noise estimation for SVM regression. Neural Networks, 17(1), 113-126.

（十四） Kalakota, R. & Robinson, M. (1999). e-Business: Roadmap for success (1st ed.). Reading, MA: Addison-Wesley.

（十五） Law, K. H. & Kwok, L. F. (2005). IDS false alarm filtering using KNN classifier. Lecture Notes in Computer Science, 3325, 114-121.

（十六） Matsumoto, R., Zhang, D., and Lu, M. (2004). Some empirical results on two spam detection methods. Proceedings of the 2004 IEEE International Conference, 198-203.