Question

我正在尝试通过对几列进行分组来找到滚动平均值。下面是我的数据集的样子：

category, sub_category,value
fruit, apple, 10
fruit, apple, 2
fruit, apple, 5
fruit, apple, 1
fruit, banana, 3
fruit, orange, 5
fruit, orange, 5
fruit, orange, 3
fruit, orange, 8

预期输出：

category, sub_category,value, rolling_average
fruit, apple, 10, 10
fruit, apple, 2, 6
fruit, apple, 5, 5.66
fruit, apple, 1, 2.66
fruit, banana, 3, 3
fruit, orange, 5, 5
fruit, orange, 5, 5
fruit, orange, 3, 4.33
fruit, orange, 8, 5.33

我能够在没有任何分组的情况下执行滚动平均，但是不确定如何在同一数据帧内按分组进行分组

Answer 1

我相信您每组需要Expanding.mean：

df['expanding_average'] = (df.groupby(['category', 'sub_category'])['value']
                             .expanding()
                             .mean()
                             .reset_index(level=[0,1], drop=True))
print (df)
  category sub_category  value  expanding_average
0    fruit        apple     10          10.000000
1    fruit        apple      2           6.000000
2    fruit        apple      5           5.666667
3    fruit        apple      1           4.500000
4    fruit       banana      3           3.000000
5    fruit       orange      5           5.000000
6    fruit       orange      5           5.000000
7    fruit       orange      3           4.333333
8    fruit       orange      8           5.250000

N=3的滚动平均值解决方案：

df['rolling_average'] = (df.groupby(['category', 'sub_category'])['value']
                           .rolling(3, min_periods=1)
                           .mean()
                           .reset_index(level=[0,1], drop=True))
print (df)

  category sub_category  value  rolling_average
0    fruit        apple     10        10.000000
1    fruit        apple      2         6.000000
2    fruit        apple      5         5.666667
3    fruit        apple      1         2.666667
4    fruit       banana      3         3.000000
5    fruit       orange      5         5.000000
6    fruit       orange      5         5.000000
7    fruit       orange      3         4.333333
8    fruit       orange      8         5.333333

熊猫-按组/子组执行滚动平均值

1 个答案: