Table 1.

Model Architecture Performance by Feature Set

	P < 0.05	Z > 2	Z > 3	Gini	Gini DT
3-Fold
Random Forest	0.717 [55.4]	0.713 [57.55]	0.714 [51.1]	0.716 [54.35]	0.717 [50.55]
KNN	0.715 [2]	0.715 [2]	0.711 [2]	0.717 [3]	0.717 [3]
SVC [linear]	0.71	0.609	0.505	0.6	0.64
SVC [poly]	0.698	0.607	0.517	0.599	0.616
SVC [sigmoid]	0.62	0.551	0.47	0.54	0.553
SVC [rbf]	0.642	0.521	0.487	0.567	0.581
Decision Tree	0.715	0.715	0.711	0.715	0.714
5-Fold
Random Forest	0.654 [41.3]	0.651 [46.75]	0.641 [45.75]	0.652 [43.95]	0.654 [41.3]
KNN	0.558 [3.39]	0.496 [9.71]	0.515 [8.44]	0.535 [3.67]	0.558 [3.39]
SVC [linear]	0.615	0.553	0.504	0.589	0.634
SVC [poly]	0.593	0.535	0.501	0.562	0.574
SVC [sigmoid]	0.551	0.489	0.458	0.517	0.521
SVC [rbf]	0.563	0.504	0.469	0.537	0.541
Decision Tree	0.634	0.636	0.634	0.64	0.637

Distribution of model accuracy using a variety of different architectures and different feature lists for both 5-fold and 3-fold cross validation methods. For KNN and Random Forest, average values for parameters with the highest accuracy are recorded in brackets