This is a collection of Python utilities that I wrote for use in machine learning.
The following are some of the most recent updates:
- added stacking_predictions_retriever.py
- to streamline my use of stacking, an ensemble learning strategy, in Kaggle competitions (see my kaggle-notebooks repository for more details), I implemented a utility that would allow me to experiment more quickly with different combinations of base models
- the utility saves predictions from the base models in files, and helps to keep track of the models that have non-stale predictions so that they can be reused, saving the time that would have otherwise been spent on making (the same) predictions using the estimators again