Model overfitting ? The training score = 1 for different count of training examples (metrics - accuracy)
Asked
Active
Viewed 378 times
-2
-
Yes, a training score of 1.0 *generally* indicates over-fitting. However, it's hard to examine how that happens for you, because you've told us nothing about the application. Are you testing with data from the training set? Is the paradigm such that a 0.95+ cross-validation is not credible? – Prune Jan 30 '17 at 17:49
1 Answers
0
Yes ,it seems like overfitting but the next thing you can do is use a test set(different from your train and cross validation set) and analyze how your model generalizes on this unseen data.

Luis Leal
- 3,388
- 5
- 26
- 49