markdown 随机森林特征重要性可视化

Posted 2021-05-06

# Ordinary Model Construction : model_rf
imp_features = pd.Series(model_rf.feature_importances_, index=X_train.columns).sort_values(ascending=False)
imp_features.plot(kind='bar', title='Feature Importance with Random Forest', figsize=(12,8))
plt.ylabel('Feature Importance Values')
plt.subplots_adjust(bottom=0.25)
plt.show()

# Pipeline Construction : pipe_rf
pipe_rf = Pipeline([('scl', StandardScaler()),
                    ('reg', RandomForestRegressor(n_estimators = 500, random_state = 10))])
pipe_rf.fit(X_train, y_train)

imp_features = pd.Series(pipe_rf.steps[1][1].feature_importances_, index=X_train.columns).sort_values(ascending=False)
imp_features.plot(kind='bar', title='Feature Importance with Random Forest', figsize=(12,8))
plt.ylabel('Feature Importance Values')
plt.subplots_adjust(bottom=0.25)
plt.show()

# ML - Random Forest Feature Importance Visualization
###### tags: `Python` `Machine Learning`

![](https://i.imgur.com/ib1y2ab.png)