Timeline for Highly unbalanced test data set and balanced training data in classification
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 16, 2018 at 17:17 | history | edited | TenaliRaman | CC BY-SA 4.0 |
deleted 8 characters in body
|
| May 2, 2018 at 14:47 | comment | added | AruniRC | Updated link to the importance estimation software from Sugiyama et al. ms.k.u-tokyo.ac.jp/software.html#uLSIF | |
| May 25, 2014 at 8:49 | comment | added | TenaliRaman | @wannik If your training and test set are random samples from the actual data, then they must have identical distributions. Almost every classifier we use expects data to be of this form. However, the situation you describe is a fairly common scenario. It is hard to predict the behaviour of the classifier in this situation. Generally, 1] Use plain classifier, if it works then great, 2] If not, do you know the class proportion in test apriori? If yes, then use transduction SVM 3] If not, then use the same approach described in the original answer (importance weights). | |
| May 25, 2014 at 7:07 | comment | added | wannik | What would happen if the training set is balanced but the testing set is not? Should they both have the same distribution? | |
| Apr 26, 2013 at 14:03 | vote | accept | user785099 | ||
| Apr 25, 2013 at 22:17 | history | answered | TenaliRaman | CC BY-SA 3.0 |