KNeighborsRegressor 的 score() 和 sklearn.metrics 的 r2_score 有什么区别

Question

print("Test set R2 score: {:.2f}".format(reg.score(X_test, y_test)))

from sklearn.metrics import r2_score print(r2_score(X_test, y_test)) enter image description here

both the methods showing different r2 score values.

pinned screenshot of the same

Answer 1

You probably meant r2_score(y_test, reg.predict(X_test)) .

sklearn.metrics.r2_score knows nothing of your model, so passing X to it is meaningless.

Answer 2

As we know in machine learning different model have different metrics.In our case it is regression model so basically for example, a least-squares regression classifier would have a score method that returns something like the sum of squared errors . The score always return the mean accuracy of the given test data and labels So this is all about score() method in knn-regressor and other regression problem as well.

The r2_score is a accuracy metrics which is calculated by the specific formula ie R2_Score=1-(RSS/TSS) where RSS--> residual sum of square and TSS--> Total sum of squares. It lies between in the range of 0 to 1 while 0 is worst and 1 is best. Rss can be calculated by squaring actual-predicted value while tss value can be calculated by squaring actual value-mean of the sample value and if u calculate these values you will get r2_score for your regression model. So basically this is the main difference for score() and r2_score()