Transportation Mode Detection Based on Permutation Entropy and Extreme Learning Machine

Table 7

Running time and accuracy of 4 : 6 data.

4 : 6	Time (s)		Training accuracy	Testing accuracy	Node	Activation function
	Training	Testing
ELM	0.0125	0.0062	49.24%	47.59%	50	Hardlim
ELM	0.0250	0.0343	55.93%	53.61%	100
ELM	0.1654	0.0780	64.05%	58.02%	300
ELM	0.4150	0.1466	69.79%	61.54%	500
ELM	1.0046	0.2153	73.05%	61.53%	800
ELM	1.5850	0.2556	74.49%	62.97%	1000
ELM	2.2683	0.2870	77.32%	63.87%	1200
ELM	3.7192	0.3796	76.40%	65.56%	800	Sigmoid
ELM	5.7836	0.5252	77.74%	65.51%	1000
ELM	8.1537	0.5564	78.31%	65.55%	1200
SVM	0.6794	0.6037	—	40.513%	—	—