With the non-store marketing more and more developed, the expansion of the logistics scale is accompanied. Due to the limited scale, the outsourcing distribution is the majority in distribution of goods.
This study focus on the performance evaluation by using data mining and value analysis. In evaluating the customer value, RFM is the most popular method used by the enterprise. However, this study adopts non-weighted RFM value in connection with the vehicle distribution characteristics which uses RFT clustering instead of the original RFM model. Rather than using the statistics average, this study concentrate on vehicle distribution characteristics (with more than one single value and the value difference) and use statistics mid-value to determine the three RFM variables.
This study use customer value matrix proposed by Marcus (1998) that throughs the clustering for dividing the vehicles into different groups by using the historical transaction information fast clustering. By using the data mining to analyze the similarity of the products which could get a better recommendation information.
The study analysis the product, vehicle, region and time period to propose a decision model, so that the manager can have a program to adjust the vehicle scheduling arrangements and which can assist the manager to have a reference to make a decision for the distribution.
