List view
Refactor the scaffold module to use an OOP interface that is more extensible and efficient than the current design. This will involve the creation of several classes. The general idea is to have: 1. a Dataset object that encapsulates common ML dataset tasks such as pre-processing and train/test splits, 2. a PandasDataset that performs these tasks for pandas DataFrame objects 3. a Model class that works with Dataset objects to predict for test sets using estimators with known interfaces. 4. a SklearnModel class that uses estimators with the scikit-learn interface. 5. a Results class that encapsulates result storage, quantitative and qualitative analysis, plotting, formatting, etc.
No due date•6/12 issues closed