Table 3.

Performance of ML algorithms for predicting ADAOO in the individuals with sporadic AD from the Paisa genetic isolate. Conventions as in Table 2. Best results are shown in bold.

ML Algorithm	Performance Measure
	RMSE		R 2		MAE
	Training	Testing	Training	Testing	Training	Testing
bstTree	3.33	5.22	0.83	0.44	2.56	3.75
glmboost	2.32	3.08	0.92	0.84	1.96	2.47
glmnet	0.25	0.52	1.00	0.99	0.17	0.39
knn	5.37	6.75	0.48	0.16	3.90	4.98
lasso	0.40	0.52	1.00	1.00	0.31	0.42
qrf	0.87	5.86	0.99	0.30	0.40	4.57
rf	2.47	5.09	0.94	0.49	1.86	4.15
rpart	5.53	7.69	0.38	0.00	4.46	6.37
rpart1SE	5.53	7.69	0.38	0.00	4.46	6.37
rpart2	5.92	6.98	0.29	0.03	4.63	5.75
svmLinear	0.61	1.11	0.99	0.97	0.57	0.83
svmLinear2	0.61	1.11	0.99	0.97	0.57	0.83
svmPoly	0.75	1.33	0.99	0.96	0.70	1.07
svmRadial	2.57	4.70	0.93	0.51	1.57	3.64
treebag	5.22	7.02	0.48	0.02	4.13	5.54
xgbLinear	0.03	4.61	1.00	0.67	0.02	3.32
xgbTree	1.13	3.98	0.98	0.70	0.93	3.19