Table 3.

Classification accuracy of $\text{NeurDNet}$ when only the first-visit tremor assessments are included in the test set.

Classifier	Binary features						Probabilistic features
	25%	35%	45%	55%	65%	75%	25%	35%	45%	55%	65%	75%
RF (entropy)	87.31	85.30	83.66	81.90	81.43	79.60	86.78	86.13	84.78	82.36	81.53	81.05
RF (gini)	87.59	85.80	83.50	82.03	80.96	79.77	86.66	85.63	84.83	82.23	81.38	80.83
SVM (rbf)	87.05	85.89	84.51	82.07	81.45	78.66	88.26	86.50	86.13	82.81	82.22	79.85
SVM (linear)	85.85	82.56	82.47	81.15	79.81	77.90	86.82	84.86	83.83	82.39	81.34	80.14
NB	84.99	83.93	79.95	81.65	77.07	75.61	87.60	86.44	85.11	84.54	82.57	81.09
LR	87.43	85.26	84.05	81.88	80.92	78.78	88.08	86.62	86.30	83.52	82.74	80.94
AdaBoost	86.26	82.53	82.70	80.21	77.69	76.46	85.79	83.59	82.32	80.78	78.63	75.06
LDA (svd)	81.13	78.10	76.10	70.13	67.04	67.82	79.12	77.02	76.49	71.58	66.14	62.99
LDA (lsqr)	81.13	78.10	76.04	70.13	62.49	49.41	79.12	77.02	76.49	71.58	66.14	51.06
QDA	79.18	80.77	77.65	70.20	62.15	60.04	93.05	89.66	77.59	71.63	59.92	54.01
DT (entropy)	80.85	79.04	78.25	76.60	76.42	74.96	79.76	79.14	78.44	77.39	75.61	73.82
DT (gini)	81.78	80.40	78.63	76.86	75.54	73.97	80.35	77.90	78.09	77.47	76.28	74.15
MLP (10)	85.41	83.33	82.54	78.90	78.09	77.60	83.31	81.84	81.50	79.72	78.15	77.83
MLP (30)	85.74	82.76	81.48	79.00	78.42	77.23	83.80	82.24	81.87	79.70	78.22	77.55

The classification accuracy is measured across different choices of second-stage classifier, including random forests (RF), support vector machines (SVM), Naive Bayes Classifier(NB), logistic regression (LR), AdaBoost classifier (AB), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), decision trees (DT), and multi layer perceptron (MLP).