rain_sizes,train_loss,test_loss=learning_curve()想要获得train_loss,test_loss,在这个函数里的scoring属性设置什么
时间: 2023-07-20 12:08:10 浏览: 46
在sklearn中的learning_curve()函数中,scoring参数用于指定模型评估的指标,但其默认值是None。由于我们想要获得train_loss和test_loss,我们需要将scoring参数设置为'mean_squared_error'或'r2',这取决于你想要使用什么评估指标。设置为'mean_squared_error'时,函数将返回训练和测试集上的均方误差;设置为'r2'时,函数将返回训练和测试集上的R方值。例如:
```
from sklearn.model_selection import learning_curve
from sklearn.linear_model import LinearRegression
train_sizes, train_scores, test_scores = learning_curve(
LinearRegression(), X, y, scoring='mean_squared_error')
train_loss = -np.mean(train_scores, axis=1)
test_loss = -np.mean(test_scores, axis=1)
```
或者
```
train_sizes, train_scores, test_scores = learning_curve(
LinearRegression(), X, y, scoring='r2')
train_loss = np.mean(train_scores, axis=1)
test_loss = np.mean(test_scores, axis=1)
```