Table 3.

Performance metric summaries comparing at tuned cross-validation in withheld samples in PPMI.

Data Modality	Genetics (P < 1E-5)	Clinico-demographic	Transcriptomics (P < 1E-2)	Combined
Stage	Tuning in PPMI	Tuning in PPMI	Tuning in PPMI	Tuning in PPMI
Algorithm	MLPClassifier	LogisticRegression	SVC	AdaBoostClassifier
AUC at training (%)	70.66	87.52	79.73	89.72
Mean, AUC during CV for baseline model (%)	69.44	88.51	78.05	86.99
Standard deviation, AUC during CV for baseline model (%)	4.46	2.17	4.27	2.30
Min, AUC during CV for baseline model (%)	62.45	86.19	71.49	84.27
Max, AUC during CV for baseline model (%)	75.73	91.98	82.62	90.70
Mean, AUC during CV for tuned model (%)	70.93	88.55	79.01	90.17
Standard deviation, AUC during CV for tuned model (%)	5.39	2.20	4.71	1.64
Min, AUC during CV for tuned model (%)	61.29	86.33	70.88	88.06
Max, AUC during CV for tuned model (%)	76.71	92.15	84.01	92.73
Variance, AUC during CV for baseline model (%)	19.89	4.73	18.20	5.29
Variance, AUC during CV for tuned model (%)	29.03	4.82	22.18	2.70