Timeline for Better understanding classification with unbalanced test data from a mathematical perspective
Current License: CC BY-SA 4.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 15, 2023 at 7:28 | comment | added | Dikran Marsupial | I posted a question asking for examples where rebalancing improves accuracy (stats.stackexchange.com/questions/559294/…) and there were no answers, even when there was a modest bonus, which suggests that this is not a real problem in practice. | |
| Apr 15, 2023 at 7:27 | comment | added | Dikran Marsupial | Note the optimal decision boundary depends on the class ratio, so if you balance the training set, the classifier is likely to over-predict the minority class, so you would have to correct for that. | |
| Apr 15, 2023 at 7:26 | comment | added | Dikran Marsupial | This all depends on how the model is constructed. If it is a generative classifier (e.g. a parametric Gaussian classifier) then the models of the distribution of patterns in each class are constructed independently, so it won't be affected by the imbalance. Similarly, if a discriminative classifier only learns about one class, it will not have minimised the cost function, so if it doesn't learn the minority class, it is in a local minima - which can't happen if the cost function is convex. So this is all classifier dependent. | |
| Apr 15, 2023 at 3:15 | answer | added | Dave | timeline score: 1 | |
| Dec 5, 2022 at 13:28 | history | edited | Dave | CC BY-SA 4.0 |
deleted 196 characters in body; edited tags
|
| Dec 5, 2022 at 13:11 | history | edited | kjetil b halvorsen♦ |
edited tags
|
|
| Dec 5, 2022 at 0:28 | comment | added | Dave | I do wonder if having a gigantic number of observations allows the signal to scream out over the noise, almost regardless of prior probability. | |
| Dec 5, 2022 at 0:13 | history | asked | Manveru | CC BY-SA 4.0 |