pandas.DataFrame.corrwith #

数据框。corrwith ( other , axis = 0 , drop = False , method = 'pearson' , numeric_only = False ) [来源] #

计算成对相关性。

成对相关性是在 DataFrame 的行或列与 Series 或 DataFrame 的行或列之间计算的。在计算相关性之前,数据帧首先沿两个轴对齐。

参数
其他数据框、系列

用于计算相关性的对象。

{0 或 'index', 1 或 'columns'}, 默认 0

要使用的轴。 0 或“索引”按行计算,1 或“列”按列计算。

drop bool, 默认 False

从结果中删除缺失的索引。

方法{'pearson', 'kendall', 'spearman'} 或可调用

相关法:

  • 皮尔逊:标准相关系数

  • kendall :Kendall Tau 相关系数

  • Spearman :斯皮尔曼等级相关

  • callable:可调用,输入两个 1d ndarray

    并返回一个浮点数。

numeric_only布尔值,默认 False

仅包含floatintboolean数据。

1.5.0 版本中的新增内容。

版本 2.0.0 中更改:默认值为numeric_onlynow False

返回
系列

成对相关性。

也可以看看

DataFrame.corr

计算列的成对相关性。

例子

>>> index = ["a", "b", "c", "d", "e"]
>>> columns = ["one", "two", "three", "four"]
>>> df1 = pd.DataFrame(np.arange(20).reshape(5, 4), index=index, columns=columns)
>>> df2 = pd.DataFrame(np.arange(16).reshape(4, 4), index=index[:4], columns=columns)
>>> df1.corrwith(df2)
one      1.0
two      1.0
three    1.0
four     1.0
dtype: float64
>>> df2.corrwith(df1, axis=1)
a    1.0
b    1.0
c    1.0
d    1.0
e    NaN
dtype: float64