jnirschl/titanic_dvc

Message	Author	SHA1	Date
Fix bug of unexpected kwarg in click.argument (only click.option has help kwarg). Update docstring	Jeff Nirschl	5c60820b25	1 year ago
Update help in click args/options. Set default input_dir relative to create_slurm_job.py. Update logging	Jeff Nirschl	f294d7c58f	1 year ago
Add newline to logging to fix output	Jeff Nirschl	c216fbd07c	1 year ago
Update input_dir default, print SBATCH settings in logging if verbose	Jeff Nirschl	88920364c6	1 year ago
Add function to create slurm job submission file	Jeff Nirschl	f16b741263	1 year ago
Add conda env specification titanic_dvc.yml	Jeff Nirschl	aedb1cc5f6	1 year ago
metrics from dvc repro in Colab	Jeff	bb4b363db6	3 years ago
Update requirements.txt to dvc ~2.1.0	Jeff Nirschl	0a9a389693	3 years ago
Merge branch 'master' of github.com:jnirschl/titanic_dvc	Jeff Nirschl	5b57868979	3 years ago
DVC repro from scratch on colab	Jeff	8f7e363090	3 years ago
Update README.md	jnirschl	768d97adc9	3 years ago
Update requirements.txt	jnirschl	1bfd9153e6	3 years ago
Fix PyYAML vulnerability in versions < 5.4.	Jeff Nirschl	6f1e8f627e	3 years ago
Update README.md to include code for getting started. Add setup.py	Jeff Nirschl	3671573669	3 years ago
set param_tuning to optimize accuracy	Jeff Nirschl	2dacbbe185	3 years ago
Update feature engineering to bin/quantize the continuous features Age, Fare, and family_size.Update parameter tuning to allow hyperopt to optimize n_estimators for RF model. Also, change hyperopt.hp.choice to hyperopt.hp.quniform for sampling integer features from a uniform distribution. Previously, hp.quniform was returning float values when the RF model required integers. The workaround included adding the following to the objective function: param[key] = int(param[key])	Jeff Nirschl	036d5e4c5f	3 years ago
Fix bug in replace_nan that removed imputation:method from params.yaml. Re-run DVC repro	Jeff Nirschl	b824d7c2d5	3 years ago
imputation method mean	Jeff Nirschl	aafaaa19bf	3 years ago
Fixed bug subfunction is_vip in build features. Re-run DVC pipeline. Re-added method-mean to params-imputation.	Jeff Nirschl	e245365204	3 years ago
Parameter tuning with hyperopt for RF model with feature engineering.	Jeff Nirschl	ad08b38ba1	3 years ago
Update script build_features.py to include 6 hand-crafted features such as family_size, is_orphan, is_single_mother etc. as well as polynomial transform features from sklearn.preprocessing. Re-run DVC repro successfully.	Jeff Nirschl	219842899f	3 years ago
exp01_rf_hyperopt-acc: parameter tuning to optimize cross-validation accuracy instead of roc_auc	Jeff Nirschl	0b75477d81	3 years ago
Set RF model in params. Re-run DVC	Jeff Nirschl	caca017d59	3 years ago
Merge pull request #5 from jnirschl/exp02_xgboost	jnirschl	7832785bbc	3 years ago
Adding back the James-Stein estimator for predicting output. RF test predictions with JS estimator = 0.77. RF test predictions without the JS estimator = 0.76.	Jeff Nirschl	846ce8b755	3 years ago
Re-run RF model predictions without the James-Stein estimator shrinkage	Jeff Nirschl	50f3aaf3e6	3 years ago
Update dvc files	Jeff Nirschl	4c4735258f	3 years ago
Add default XGBoostClassifier. Re-run pipeline	Jeff Nirschl	17d9608490	3 years ago
Re-do DVC stage predict_output. I had forgotten to add models/estimator.pkl as a dependency for the stage. Running successfully now.	Jeff Nirschl	186090e400	3 years ago
Add DVC stage predict_output. Current best submission with RF model + hyperopt and predicting proba with JS estimator gives a test-set accuracy of 0.77033, which is not too far from the cross-validation accuracy.	Jeff Nirschl	1a0beb71bf	3 years ago

Newer Older

jnirschl / titanic_dvc mirror of https://github.com/jnirschl/titanic_dvc.git

jnirschl
/
titanic_dvc
mirror of https://github.com/jnirschl/titanic_dvc.git