Transportation Mode Detection Based on Permutation Entropy and Extreme Learning Machine

Table 5

Running time and accuracy of 2 : 8 data.

2 : 8	Time (s)		Training accuracy	Testing accuracy	Node	Activation function
	Training	Testing
ELM	0.0125	0.0125	48.16%	46.42%	50	Hardlim
ELM	0.0125	0.0312	58.28%	55.66%	100
ELM	0.0842	0.0998	64.90%	57.81%	300
ELM	0.2496	0.1716	71.26%	61.58%	500
ELM	0.6739	0.2621	74.79%	62.19%	800
ELM	1.0265	0.3245	77.16%	64.06%	1000
ELM	1.6068	0.3713	77.83%	62.70%	1200
ELM	1.9438	0.4742	76.34%	64.28%	800	Sigmoid
ELM	3.1013	0.5959	77.94%	63.45%	1000
ELM	4.5053	0.7582	80.16%	65.48%	1200
SVM	0.1723	0.3869	—	40.068%	—	—