2
votes
Model Performance on external validation Set really low?
First, an AUC less than 50% is terrible: it means that you get better performance by switching the positive and negative labels! So the model is doing worse than nothing on this data.
In general there ...
- 24.5k
2
votes
Is my model overfitting ? Training Acc :93 % test accuracy 82%
How can we ask this question? Results doesn't simply depends on 'what is the accuracy on training and test sets', there are others side to consider.
What about your data? Is it balanced or not? What ...
- 68
2
votes
What is the proper way to use early stopping with cross-validation?
I once wondered the same in case of LightGBM and got this answer from its creator, Guolin Ke:
I think in both XGBoost and LightGBM, the CV will use the average
scores from all folds, and use this for ...
- 123
2
votes
Accepted
Online Learning/Continual Learning for tree-based Algorithms
This is a really good question for which I will give you a theoretical result; in particular, I am not aware of any specific implementation in any programming language.
The concept of incremental ...
- 639
2
votes
Optuna Median Pruner n_warmup_steps
The steps in n_warmup_steps refer to the incremental steps while gradient decent. So with ...
- 318
2
votes
Accepted
Why is monotonic constraint disabled when using MAE as an objective to LGBM?
MAE is an odd loss function for GBMs: the gradient is constant ($\pm1$), and the hessian all zeros, so the usual tree-training target of $-G/H$ (possibly with additional terms for regularization) ...
- 10.8k
1
vote
Combining results from classifiers trained on different test/train splits results in higher accuracy
A number of observations as per your use case:
Please use a k-fold validation scheme, for more reliable numbers of accuracy.
The model or data need some more hyper-parameter tuning since ...
1
vote
Accepted
Grid-search for a multi-output regression task using Scikit-learn's API
I assume you meant GridSearchCV(estimator ..., otherwise there's no wrapping here.
You'll need to supply a prefix:
...
- 656
1
vote
Accepted
How to improve Regression RMSE with LightGBM
mean_squared_error(y_pred,y_test) is MSE, not RMSE (which would be mse ** 0.5). Taking a square root of it yields around 80k, ...
- 656
1
vote
How to make LightGBM to suppress output?
I read all the answers and issues, and tried all these approaches and yet LGBM still outputs some info (which drives me crazy). If you want to completely suppress any output during the training try ...
Only top scored, non community-wiki answers of a minimum length are eligible