Machine Learning Prep and Train Toolkit

The ML Prep and Train project originated to address the need for a standardized approach to preparing datasets and training machine learning models. The toolkit provides functionality for cleaning and structuring data, feature extraction, and utilizing various machine learning algorithms for binary classification. With an emphasis on efficiency and best practices, this project serves as a cornerstone for data scientists needing to quickly iterate and test hypotheses in their binary classification problems.

Training does not stop at fitting one algorithm. Each candidate gets its own hyperparameter search, optional probability calibration, and an evaluation table reporting accuracy, precision, recall, F1, AUC and Gini beside an overfitting probability computed both by accuracy and by F1 — so a model that only looks strong on the data it was trained on is flagged rather than quietly shipped. The base classifiers then feed a stacking model, with a logistic regression, a gradient-boosted tree, or a small neural network sitting on top of their outputs.

Because a binary classifier is nearly always asked to justify itself, every model produces an explanation alongside its score: a feature-importance plot where the algorithm exposes one, and a LIME explanation where it does not.

Feel free to explore the project's repository on GitHub and contribute or adapt the toolkit to your specific needs.