To maximize interpretability of model differences, we must have the same folds for all SVM models and NN models.
[x] Programmatically divvy accounts into 10 folds and record assignment to be called when SVM classifiers are called
[x] Use assignment to re-parse NN datasets into 10 folders (1 per fold)
[x] OverweightClassifier, OverweightSuite, and OverweightEnsemble command-line option to use US overweight proportion (~70%) OR collected proportion (~12%).
To maximize interpretability of model differences, we must have the same folds for all SVM models and NN models.
OverweightClassifier,OverweightSuite, andOverweightEnsemblecommand-line option to use US overweight proportion (~70%) OR collected proportion (~12%).