如何使用matplotlib定义要素重要性得分中从最大值到最小值的顺序？

Question

我为我的数据集计算重要性分数，数据集包含30个特征和假设类的I列。而且确实有阴谋。但数组中的值不是按特征重要性得分排序的。numpy或matplotlib中类的开关变量由它管理

from pandas import read_csv, read_excel

from sklearn.ensemble import ExtraTreesClassifier

# load data


names = ['8,oct18', '3,oct18', '4,oct18', '3,sen17', '3,sen17', '4,sen17', '8,sen17', '3,aug17', 
         '8,aug17','4,aug17', '3,apr17', '4,apr17', '8,apr17', '3,jan17', '8,jan17', '4,jan17', 'jan19', 'jan19',
        'jan19', 'may18', 'may18', 'may18', '11, sen17', '11, dec2017', '12,dec2017', '11,aug 2017', 
         '12,aug 2017', '11, apr 2017', '12, apr 2017', '30t', 'class']

dataframe = read_excel("/home/qw/myprojects/valuevo/data.xlsx", names = names)

array = dataframe.values

X = array[:,0:30]

Y = array[:,30]

# feature extraction

model = ExtraTreesClassifier()

model.fit(X, Y)

print(names, '=', model.feature_importances_)

import matplotlib.pyplot as plt
import numpy as np
#case x,y were 31x30 before this:
names = ['8,oct18', '3,oct18', '4,oct18', '3,sen17', '3,sen17', '4,sen17', '8,sen17', '3,aug17', 
         '8,aug17','4,aug17', '3,apr17', '4,apr17', '8,apr17', '3,jan17', '8,jan17', '4,jan17', 'jan19', 'jan19',
        'jan19', 'may18', 'may18', 'may18', '11, sen17', '11, dec2017', '12,dec2017', '11,aug 2017', 
         '12,aug 2017', '11, apr 2017', '12, apr 2017', '30t']
x = np.array(names)
y = np.array(model.feature_importances_)
plt.title("RF score") 

plt.plot(y,x)

接下来是一个情节。如何按值排序分数？

Answer 1

您可以使用np.argsort获得值的排序顺序，并使用该索引将x和y值一起排序：

sort_ix = np.argsort(model.feature_importances_)
x = np.array(names)[sort_ix]
y = np.array(model.feature_importances_)[sort_ix]
plt.plot(y, x)

我不知道为什么要调用x值y和y值x。这有点令人困惑