Removing the “bad” samples will for sure improve the performance in the training set, but how about the “bad” samples in the test set and in unseen samples when you use the perfect model to make predictions?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| For tip4 - toss out bad example | 1 | 56 | May 18, 2023 | |
| Tossing out bad examples: Real world production data distribution | 8 | 421 | August 11, 2021 | |
| How Do We Best Train For Noisy Data? (Can A Staged Training Strategy for Noisy Data Improve Results?) | 1 | 54 | May 16, 2023 | |
| "Bad" Data For More Robust Algorithm | 1 | 58 | May 16, 2023 | |
| Week2 doubt- trying out model on subset of data | 1 | 548 | June 10, 2021 |