python - 对 Pandas 数据框选定列中的每一行应用 zscore 函数

Question

使用以下代码:

import pandas as pd 
from sklearn.preprocessing import scale

df = pd.DataFrame({"Probe":["1430378_at","1439896_at","1439896_at"],
                   "Gene":["2900011G08Rik","Trappc5","Limk2"],
                   "A.x1":[0.0767,  0.4383,  0.7866],
                   "A.x2":[0.8091,  0.1954,  0.6307],
                   "A.x3":[ 0.6599,  0.1065,  0.0508]
                  }
                  )
df = df[["Probe","Gene","A.x1","A.x2","A.x3"]]

我可以获得以下数据框:

In [55]: df
Out[55]:
        Probe           Gene    A.x1    A.x2    A.x3
0  1430378_at  2900011G08Rik  0.0767  0.8091  0.6599
1  1439896_at        Trappc5  0.4383  0.1954  0.1065
2  1439896_at          Limk2  0.7866  0.6307  0.0508

我想做的是计算 Ax1,x2,x3 列的 z-score。我怎样才能做到这一点？

例如，对于第二行，我们使用以下方法计算 z-score

from sklearn.preprocessing import scale
scale([0.4383,0.1954,0.1065],axis=0,with_mean=True, with_std=True,copy=False)
Out[61]: array([ 1.36603199, -0.36604999, -0.999982  ])

最终我们期望得到:

        Probe           Gene        A.x1    A.x2        A.x3
0  1430378_at  2900011G08Rik  -1.38769528  0.92991195  0.45778333
1  1439896_at        Trappc5  1.36603199  -0.36604999  -0.999982
2  1439896_at          Limk2  0.93889666  0.44644183  -1.3853385

Answer 1

对于 Pandas，尝试将 apply 与匿名函数一起使用来对每一行执行计算通常是个好主意。这对你有用吗？:

    df.iloc[:,2:5] = df.filter(regex = 'A.x').apply(
            lambda V: scale(V,axis=0,with_mean=True, with_std=True,copy=False),axis=1)