scipy.stats具有函数^{},该函数允许您计算一个值高于平均值的标准差(通常称为标准分数或Z分数)。在
如果arr是您问题中的示例数组,那么您可以计算每行25的Z分数,如下所示:>>> import scipy.stats as stats
>>> stats.zscore(arr, axis=1)
array([[-0.18017365, -0.52666143, -0.52666143, 0.8592897 , -0.52666143,
-0.18017365, -0.52666143, -0.52666143, -0.52666143, -0.52666143,
0.51280192, -0.52666143, 0.16631414, -0.18017365, -0.52666143,
-0.52666143, 0.51280192, -0.52666143, 0.51280192, -0.52666143,
4.32416754, -0.52666143, 0.16631414, -0.52666143, 0.8592897 ],
[-0.43643578, -0.43643578, -0.43643578, 0.47280543, 1.38204664,
-0.43643578, -0.43643578, -0.43643578, -0.43643578, -0.43643578,
0.47280543, -0.43643578, 1.38204664, -0.43643578, -0.43643578,
-0.43643578, -0.43643578, -0.43643578, -0.43643578, -0.43643578,
4.10977027, -0.43643578, -0.43643578, -0.43643578, 0.47280543]])
此计算使用每行的总体平均值和标准差。要使用样本方差(与t统计量一样),另外指定ddof=1:
^{pr2}$