如何使用 pandas 找到每个主成分的前三个特征？

Question

我正在按照给定的解决方案。

但是解决方案从每个主成分中获取 argmax() 特征。我想拿前三名。我该怎么做？

我基本上想分别知道哪些功能对每台 PC 的影响最大。

谢谢。

Answer 1

您可以使用np.argsort 或np.argpartition 获取排序后的索引。按照指示的问题程序

# With argsort 
most_important = [np.argsort(np.abs(model.components_[i]))[::-1][:3] for i in range(n_pcs)]

# With argpartition
most_important = [np.argpartition(np.abs(model.components_[i]), -3)[-3:] for i in range(n_pcs)]

most_important
>>> [array([4, 1, 0]), array([2, 3, 4])]

然后将最重要的组件作为列

initial_feature_names = ['a','b','c','d','e']

# Notices the [::-1] is used to order the component names
most_important_names = [[initial_feature_names[i] for i in most_important[i][::-1]] for i in range(n_pcs)]
dic = {'PC{}'.format(i): most_important_names[i] for i in range(n_pcs)}
pd.DataFrame.from_dict(dic).T
>>>
    0   1   2
PC0 e   b   a
PC1 c   d   e

如何使用 pandas 找到每个主成分的前三个特征？

How to find the top three features of every principal component using pandas?

python

pca

pandas