this repo has no description
feat: 📝 SHAP
+3
-2
+3
-2
Datathons.md
+3
-2
Datathons.md
···
7
7
3. Get baseline model working.
8
8
4. Create `scikit-learn` compatible metric if needed.
9
9
5. Try different approaches for preprocessing (encodings, Deep Feature Synthesis, lags, aggregations, imputers, ...)
10
-
- If yopu're working as a group, split preprocessing feature generation between files.
10
+
- If yopu're working as a group, split preprocessing feature generation between files.
11
11
6. Plot learning curves ([sklearn](https://scikit-learn.org/stable/modules/learning_curve.html) or [external tools](https://github.com/reiinakano/scikit-plot)) to avoid overfitting.
12
12
7. Tune hyper-parameters once you've settled on an specific approach. ([optuna](https://optuna.readthedocs.io/)).
13
13
8. Plot and visualize the predictions (histograms, random prediction, ...) to make sure they're doing as expected.
14
-
9. Think about what postprocessing heuristics can be done to improve or correct predictions.
14
+
- Explain the predictions with [SHAP](https://github.com/slundberg/shap)
15
+
9. Think about what postprocessing heuristics can be done to improve or correct predictions.
15
16
10. [Stack](https://scikit-learn.org/stable/auto_examples/ensemble/plot_stack_predictors.html) classifiers ([example](https://www.kaggle.com/couyang/featuretools-sklearn-pipeline#ML-Pipeline)).
16
17
11. Try AutoML models:
17
18
- For tabular data: [TPOT](https://github.com/EpistasisLab/tpot), [AutoSklearn](https://github.com/automl/auto-sklearn), [AutoGluon](https://auto.gluon.ai/stable/index.html), Google AI Platform, [PyCaret](https://github.com/pycaret/pycaret), [Fast.ai](https://docs.fast.ai/), [Alex](https://github.com/Alex-Lekov/AutoML_Alex).