pandas GroupBy并计算Z分数 [英] Pandas GroupBy and Calculate Z-Score

问题描述

所以我有一个看起来像这样的数据框:

So I have a dataframe that looks like this:

pd.DataFrame([[1, 10, 14], [1, 12, 14], [1, 20, 12], [1, 25, 12], [2, 18, 12], [2, 30, 14], [2, 4, 12], [2, 10, 14]], columns = ['A', 'B', 'C'])

    A   B   C
0   1   10  14
1   1   12  14
2   1   20  12
3   1   25  12
4   2   18  12
5   2   30  14
6   2   4   12
7   2   10  14

我的目标是获取B列相对于A列和C列的z分数.我知道我可以计算每组的均值和标准差

My goal is to get the z-scores of column B, relative to their groups by column A and C. I know I can calculate the mean and standard deviation of each group

test.groupby(['A', 'C']).mean()    
        B
A   C   
1   12  22.5
    14  11.0
2   12  11.0
    14  20.0

test.groupby(['A', 'C']).std()
        B
A   C   
1   12  3.535534
    14  1.414214
2   12  9.899495
    14  14.142136

现在，我要根据这些均值和标准差来计算B列中的每个项目的z得分.因此，第一个结果将是(10-11)/1.41.我觉得必须有一种无需太多复杂性就能做到这一点的方法，但是我一直坚持如何进行.让我知道是否有人可以向我指出正确的方向，或者我是否需要澄清任何事情！

Now for every item in column B I want to calculate it's z-score based off of these means and standard deviations. So the first result would be (10 - 11) / 1.41. I feel like there has to be a way to do this without too much complexity but I've been stuck on how to proceed. Let me know if anyone can point me in the right direction or if I need to clarify anything!

推荐答案

使用transform

Mean=test.groupby(['A', 'C']).B.transform('mean')    
Std=test.groupby(['A', 'C']).B.transform('std')

然后

(test.B - Mean) / Std

---

scipy

from scipy.stats import zscore
test.groupby(['A', 'C']).B.transform(lambda x : zscore(x,ddof=1))
Out[140]: 
0   -0.707107
1    0.707107
2   -0.707107
3    0.707107
4    0.707107
5    0.707107
6   -0.707107
7   -0.707107
Name: B, dtype: float64

这篇关于 pandas GroupBy并计算Z分数的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！