-
There are several evaluation metrics that would be particularly beneficial for (binary) imbalanced classification problems and would be greatly appreciated additions. In terms of prioritizing implemen…
-
But with multi class,it doesnt works
by e
class 1 = 2000 trained cases
class 2 = 10 trained cases
class 3 = 400 trained cases
class 4 = 160 trained cases (case: text blue house)
pipeli…
-
The Glass + CH data has highly imbalanced SIC code classifications. The distribution of SIC Sections is shown below and this is more pronounced at the 4-digit level (some codes only have one example).…
-
Hi,
@albahnsen Do you have any thoughts on the relationship between the cost matrix and re-balancing the data? I notice that you do not rebalance in your final logistic regression model in your wo…
S-C-H updated
4 years ago
-
Test Focal Loss:
Label distribution:
0 - 410
1 - 3090
2 - 4037
3 - 3259
4 - 4197
-
# Add more analytics and graphs for economy monitoring
## Context
Users have expressed interest in having more detailed analytics and graphs to monitor the in-game economy. This will help server owne…
-
### Describe the issue linked to the documentation
Hi guys,
In the "ROC curve using micro-averaged OvR" part of the doc (https://scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html#r…
-
Creating an issue to track misc balancing issues we want to keep an eye on or address:
## Issues
- [ ] Stalwarts don't feel good -- their shield buff doesn't defend against light / melee units eno…
-
There's probably no perfect way to do this, but many Clojure/lisp environments (e.g. Cursive, Counterclockwise, some Emacs setups, etc.) do a decent job, seemingly by going top to bottom and re-format…
-
Hello authors,
I’ve noticed that most of the time spent running `ray_processing/tokenize_shuffle.py` is during the buffer write process, while the tokenization itself is relatively fast. Do you hav…