結合田口方法及倒傳遞類神經網路建構連鎖超商選址評估模式之研究

• Main Authors: Cheng-Chang Tsai, 蔡政璋
• Other Authors: Hsin-Pin Fu
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/04547679762476065146

碩士 === 國立高雄第一科技大學 === 行銷與流通管理研究所 === 104 === With the fierce competition of the convenience stores in Taiwan, business district and site condition are keys to the decision of opening branches. With decades of business experience, Taiwan’s well-known convenience chain stores have built their own strategic assessments and systematic models on location selection. Many past researches have studied effective factors of convenience stores’ location selection; yet, none de-rived find-ings thru the connection with the actual performance, turnover for example. Actual business performance, however, provides an important indication and objective assessment for the store opening decision. This study used data collected from a Taiwan well-known convenience chain store, including stores’ types of business district, competitors in the district, people passage, sources of customers, numbers of residential hous-es, visual condition of the location, store inside area, actual monthly turno-ver, and adopted Back-Propagation Network (BPN) combined with Taguchi experimental research method to develop a predicting model for store opening decision. The analytical results indicate the best artificial neural network, a two-layer network with 3000 network learning iterations, 0.7 of learning rate, 0.4 of momentum term and hyperbolic tangent as data trans-fer function, to predict a store’s turnover based on the inputs of its lo-cation condition is able to make 76% of prediction rate.