sklearn笔记

代码篇

GridSearchCV

sklearn.model_selection.GridSearchCV

class sklearn.model_selection.GridSearchCV(estimator, param_grid, scoring=None, n_jobs=None, iid=’warn’, refit=True, cv=’warn’, verbose=0, pre_dispatch=‘2*n_jobs’, error_score=’raise-deprecating’, return_train_score=False)

Exhaustive search over specified parameter values for an estimator.

Important members are fit, predict.

GridSearchCV implements a “fit” and a “score” method. It also implements “predict”, “predict_proba”, “decision_function”, “transform” and “inverse_transform” if they are implemented in the estimator used.

The parameters of the estimator used to apply these methods are optimized by cross-validated grid-search over a parameter grid.

返回一个model，等待训练，示例如下：

#Finding best parameters for our SVC model
'''
svm的参数，
'''
param = {
    'C': [0.1,0.8,0.9,1,1.1,1.2,1.3,1.4],   
    'kernel':['linear', 'rbf'],
    'gamma' :[0.1,0.8,0.9,1,1.1,1.2,1.3,1.4]
}
grid_svc = GridSearchCV(svc, param_grid=param, scoring='accuracy', cv=10)  # 10折交叉验证，返回模型

训练模型：

grid_svc.fit(X_train, y_train) 
#Best parameters for our svc model
grid_svc.best_params_   # 打印最好的一组参数
grid_svc.best_score_    # Mean cross-validated score of the best_estimator

# 找到最好的一组参数后，使用最好的参数去训练模型
svc2 = SVC(C = 1.2, gamma =  0.9, kernel= 'rbf')
svc2.fit(X_train, y_train)
pred_svc2 = svc2.predict(X_test)
# Build a text report showing the main classification metrics
print(classification_report(y_test, pred_svc2)) 
print(confusion_matrix(y_test, pred_svc2))

理论篇

regression和classification里的损失函数选择

一般地，regression => Cross-Entropy , classification => Square Error

正则化参数 $\alpha$ 的选取

对于正则化项：$\alpha ||W|{|}_p^p$ 一般常使用GridSearchCV来搜寻最佳参数。

机器学习常用的评价指标

链接

学习中遇到的分类任务中的评价指标有准确率（Accuracy）、FPR、FNR、Recall、Precision、F-score、MAP、ROC曲线和AUC等，回归任务中的指标有**®MSE、MAE、CC/PCC**等。

• $R^2$
  
• MSE / RMSE / MAE