Work
Python Library
UAutoml
An automated machine learning framework for data analysis, feature selection, multi-model training, grid-search tuning, and cross-validated evaluation, packaged on PyPI.
01Problem
Automate the repetitive parts of model selection.
The framework handles datasets with both numerical and categorical features and automates feature selection, model training, hyperparameter tuning, and evaluation for classification and regression tasks.
02Approach
Compare models, tune with grid search, validate.
It trains decision trees, SVMs, logistic regression, random forests, and multi-layer perceptrons, automatically optimizes hyperparameters via grid search, and uses cross-validation (accuracy for classification, MSE for regression) to surface the best classification and regression model.
The whole thing is distributed as a reusable Python library on PyPI.