Table 3.

Evaluation of machine learning approaches on weather and air quality.

Machine learning approaches		F1 measure	Accuracy, % (n/N)
Generalized linear model			85.6 (479/559)
	Peak	0.667
	Nonpeak	0.908
Support vector machine			80.2 (448/559)
	Peak	0.289
	Nonpeak	0.882
Adaptive boosting neural networks			84.7 (473/559)
	Peak	0.667
	Nonpeak	0.900
Tree bag			83.8 (468/559)
	Peak	0.640
	Nonpeak	0.895
Random forest			88.3 (494/559)
	Peak	0.745
	Nonpeak	0.924