I was wondering if the intersection between the two Train Error lines had any meaning or not? Is there a reason why the train error on 5000 started with a greater error than the one with 100 000 samples?
I'm just speculating but maybe it could be that given a larger training set, the data could be more balanced (if that makes sense) so some initial model configuration would have a lower error than a smaller dataset which could be more skewed due to its size. Just an idea but I would also like to know if this is intentional or not :)
Exam 2017 Problem 22
Hello,
I was wondering if the intersection between the two Train Error lines had any meaning or not? Is there a reason why the train error on 5000 started with a greater error than the one with 100 000 samples?
3
I'm just speculating but maybe it could be that given a larger training set, the data could be more balanced (if that makes sense) so some initial model configuration would have a lower error than a smaller dataset which could be more skewed due to its size. Just an idea but I would also like to know if this is intentional or not :)
2
Add comment