如何从特征列表和权重系数列表中 select 前 10 个特征（逻辑回归）？

Question

我想 select 我的 Logistic 回归模型中的前 5 个特征。我现在有两个数组，一个包含所有特征名称，另一个列表包含来自 model.coef_ 的系数，其中模型 = LogisticRegression()。

feature_list = ['ball', 'cat', 'apple',....,] # this has 108 elements
coefficents = lr.coef_  
print(coefficents[0])

打印如下：

[ 2.07587361e-04  5.59531750e-04  0.00000000e+00  0.00000000e+00
-5.16353886e-02 ......  1.66633057e-02]   #this also has 108 elements

当我尝试对 coeff 值进行排序时，我得到了不同的值。

sorted_index = np.argsort(coefficents[0])
print(sorted_index)
[ 22  91  42  15  52  31  16  32  86 .... 17 106]   #this has 108 values

如何从这两个数组中获取正确的前 5 个重要特征？

Answer 1

argsort是升序排列，你要降序（从大到小）

这里我举个简单的例子：

import numpy as np

feature_list = ['ball', 'cat', 'apple', 'house', 'tree', 'school', 'child']
coeff = np.array([0.7, 0.3, 0.8, 0.2, 0.4, 0.1, 0.9])
# negate the coeff. to sort them in descending order
idx = (-coeff).argsort()
# map index to feature list
desc_feature = [feature_list[i] for i in idx]
# select the top 5 feature
top_feature = desc_feature [:5]
print(top_feature)

结果是您的主要特征：

['child', 'apple', 'ball', 'tree', 'cat']

如何从特征列表和权重系数列表中 select 前 10 个特征（逻辑回归）？

How to select top 10 features from a list of features and a list of weight coefficients ( Logistic Regression )?

machine-learning

feature-selection

python-3.x

logistic-regression