Price Response Medel Analysis on Daily POS Data

我们建立了2000年牛奶类和2001年其他类的每日POS数据库，并进行了分析。2002年，我们以“洋品牌与国产品牌”、“自有品牌与民族品牌”为专题进行了比较。数据库的构建：除了之前购买的1994年和1995年的POS数据外，还购买了1996年至2000年全国100多家门店的POS数据。这是一个巨大的数据文件量，我们制作了程序，使时间序列的任何产品类别的日常POS数据。此外，我们还建立了关于“商店类型”、”商店地址的本地特征”和“有无停车场”等商店特征的附加数据库。POS数据分析：牛奶类别：对牛奶类别的POS数据特征进行了研究。结果发现：价格反应函数（S曲线）的非线性，删失数据（数据分布截止），存在星期效应，源自公司决定货架上有多少件商品。另一类：重复购买的国产品牌与只购买的国外品牌在促销效果的时滞上存在差异，促销活动在速溶咖啡、薯片、黄油、人造黄油、意大利面和蛋黄酱这几个品类上，我们对“自有品牌”和“国产品牌”进行了比较分析，对于后两者，我们采用了联立方程的方法。

We constructed the database of daily POS data for milk category in 2000 and the other categories in 2001 and the analyzes were conducted. In 2002, we conducted the comparison of "foreign brand vs. domestic brand" and "private brand vs. national brand" as particular themes.Construction of Database : We purchased POS data of more than 100 stores in nationwide for the period of 1996 to 2000 in addition to those of 1994 and 1995 purchased before. This is a huge volume of data file and we made programs to make time series for any product category of daily POS data. Further, we made additional database on the characteristics of store such as "store type"," local characteristics of store addresses", and "the existence of paring lot".The analyzes on POS data :Milk category : The characteristics of POS data were investigated on milk category. As a result, the following was observed : the non-linearity of price response function(S-curve), censored data (the data distribution cut-off), the existence of day-of-week effect derived from the company decision on how many units of an item are stocked on the shelf.The other category : there was a difference on the time delay of promotional effect between domestic brands that are repeatedly purchased and foreign brands that are purchased only on promotional occasion. On the categories of instant coffee, potato chips, butter, margarine, pasta and mayonnaise, we made analyzes on the comparison of "private brand" and "national band", and for the latter two, we used simultaneous equations.