Contents
Why is01 loss non-convex?
0–1 loss is robust to out- liers since it is not affected by a misclassified point’s distance from the margin, but this property also makes it non-convex; the convex squared, hinge, and log losses are not robust to outliers in this way since their penalty does scale with the margin of misclassification.
Why use01 loss function?
3 Answers. You have correctly summarized the 0-1 loss function as effectively looking at accuracy. In this case, your loss function would weight false negative misclassification more heavily. The sum of your losses would no longer represent accuracy in this case, but rather the total “cost” of misclassification.
Why is 0-1 loss not used frequently?
The 0-1 loss function is non-convex and discontinuous, so (sub)gradient methods cannot be applied.
How to explain the 0-1 loss function?
If you will try ti explain the concept using different example, please use 0-1 loss function. You have correctly summarized the 0-1 loss function as effectively looking at accuracy. Your 1’s become indicators for misclassified items, regardless of how they were misclassified.
What is the accuracy of the loss function?
Since you have three 1’s out of 10 items, your classification accuracy is 70%. If you change the weighting on the loss function, this interpretation doesn’t apply anymore.
What happens when you change the loss function?
Your 1’s become indicators for misclassified items, regardless of how they were misclassified. Since you have three 1’s out of 10 items, your classification accuracy is 70%. If you change the weighting on the loss function, this interpretation doesn’t apply anymore.
Which is the loss for one data point?
Specifically, your L ( y, y ^) is the loss for one data point (I am changing the notation little bit). And the loss for the whole data set, i.e., classification accuracy, needs to summing all data points. Thanks for contributing an answer to Cross Validated!