本文共 1223 字，大约阅读时间需要 4 分钟。

# Show your work here - the plot below was helpful for me

# https://stackoverflow.com/questions/44101458/random-forest-feature-importance-chart-using-python features = diabetes.columns[:diabetes.shape[1]] print("features=",features) importances = random_search.best_estimator_.feature_importances_ print("importances=",importances) indicies = np.argsort(importances) print("indicies=",indicies) print("importances[indicies]=",importances[indicies]) plt.figure(1) plt.barh(range(len(indicies)),importances[indicies],color='b',align='center') plt.yticks(range(len(indicies)),features[indicies]) plt.xlabel("realative correlation") plt.figure(2) plt.bar(range(len(indicies)),importances[indicies]) plt.xticks(range(len(indicies)),features[indicies],rotation=45)

#效果

features= Index(['Pregnancies', 'Glucose', 'BloodPressure', 'SkinThickness', 'Insulin',       'BMI', 'DiabetesPedigreeFunction', 'Age', 'Outcome'],      dtype='object')importances= [ 0.04620102  0.35829186  0.06762202  0.04479566  0.0593829   0.17193048  0.10250037  0.14927569]indicies= [3 0 4 2 6 7 5 1]importances[indicies]= [ 0.04479566  0.04620102  0.0593829   0.06762202  0.10250037  0.14927569  0.17193048  0.35829186]

下面两图的意义是各个特征值与输出标签的相关系数，用横纵柱状图进行描绘

 

                            

转载地址：http://lsfti.baihongyu.com/