Transportation Mode Detection Based on Permutation Entropy and Extreme Learning Machine

Table 8

Running time and accuracy of 5 : 5 data.

5 : 5	Time (s)		Training accuracy	Testing accuracy	Node	Activation function
	Training	Testing
ELM	0.0229	0.0109	49.16%	49.22%	50	Hardlim
ELM	0.0504	0.0268	54.90%	54.89%	100
ELM	0.2454	0.0755	64.69%	62.39%	300
ELM	0.5117	0.1309	68.09%	63.49%	500
ELM	1.1518	0.1908	71.87%	65.31%	800
ELM	1.7061	0.2378	73.55%	65.61%	1000
ELM	3.1928	0.2912	74.70%	65.87%	1200
ELM	5.3716	0.4472	74.03%	67.90%	800	Sigmoid
ELM	8.0185	0.4628	75.74%	69.03%	1000
ELM	11.6845	0.6240	77.55%	69.27%	1200
SVM	2.382393	1.171633	—	42.816%	—	—