数据包括,产品分类(颜色和材料)product (I dummy them),折扣方式sales (Special sale, friends&family, regular website sales, no sale),日期(month),售出量qty,单价price。
然后我就
ls qty c price @expand(sales, @droplast)
Dependent Variable: QTY |
|
| ||
Method: Least Squares |
|
| ||
Date: 11/16/09 |
|
| ||
Sample: 1 1759 |
|
|
| |
Included observations: 1759 |
|
| ||
|
|
|
|
|
|
|
|
|
|
Variable | Coefficient | Std. Error | t-Statistic | Prob. |
|
|
|
|
|
|
|
|
|
|
C | 20.36327 | 1.535365 | 13.26282 | 0.0000 |
PRICE | -0.517092 | 0.044562 | -11.60386 | 0.0000 |
SALES=FRD&FML | 4.199580 | 0.527503 | 7.961251 | 0.0000 |
SALES=REG | 16.90276 | 1.602402 | 10.54839 | 0.0000 |
SALES=REG_SALE | 9.241078 | 0.967490 | 9.551597 | 0.0000 |
|
|
|
|
|
|
|
|
|
|
R-squared | 0.103090 | Mean dependent var | 2.001706 | |
Adjusted R-squared | 0.101045 | S.D. dependent var | 3.727263 | |
S.E. of regression | 3.533939 | Akaike info criterion | 5.365542 | |
Sum squared resid | 21905.22 | Schwarz criterion | 5.381097 | |
Log likelihood | -4713.994 | Hannan-Quinn criter. | 5.371291 | |
F-statistic | 50.40095 | Durbin-Watson stat | 1.345502 | |
Prob(F-statistic) | 0.000000 |
|
|
|
|
|
|
|
|
|
|
|
|
|
这也太恶心了!基本上不成立啊!
我也试过了别的东西但是基本上那个product根本没意义。价格和特价是最重要的。
问题是怎么样才能让这model好看点?Lag?LOG?。。
HELP!!!


雷达卡



京公网安备 11010802022788号







