Table 2.

Final optimized classifier parameters, down-sampling rate, and reduction method.

Concept	Kernel	Parameter	Reduction	Downsample	Sentence Validation + testing
					AUC	AP
Autonomic_dysfunction	rbf	0.01	max	0.05	0.68	0.01
Cerebral_palsy	Linear	0.1	noisy_or	1.0	0.93	0.62
Developmental_delay	Linear	0.1	noisy_or	0.95	0.96	0.59
Epilepsy_or_seizures	Linear	0.01	noisy_or	1.0	0.94	0.57
Feeding_issues	Linear	0.01	noisy_or	0.90	0.93	0.33
Hypotonia	Linear	0.01	noisy_or	0.60	0.94	0.36
Insomnia	rbf	0.0001	noisy_or	0.60	0.90	0.02
Mood_disturbance	Linear	0.005	noisy_or	1.0	0.95	0.23
Movement_disorders	Linear	0.1	max	0.75	0.74	0.002
Oculogyric_crisis	Linear	8.0	max	1.00	0.93	0.002

Sentence level performance on the validation + testing datasets shown for AUC (area under the receiver operating curve) and AP (average precision).