Novel FTLRNN with Gamma Memory for Short-Term and Long-Term Predictions of Chaotic Time Series

Table 19

For K = 12 , training and testing samples variation for FTLRNN on testing data set for monthly sunspot time series and K = 10 for laser time series.

Time series		Sunspot time series 𝐾=12			Laser time series 𝐾=10
Training exemplars	Testing exemplars	MSE	NMSE	𝑟	MSE	NMSE	𝑟
10%	75%	0.01187	0.32145	0.87483	0.01506	0.45940	0.76210
20%	65%	0.01116	0.29307	0.84917	0.01310	0.38332	0.80583
30%	55%	0.00967	0.23485	0.88221	0.00140	0.43507	0.77113
40%	45%	0.00749	0.17299	0.91236	0.01221	0.48119	0.75879
50%	35%	0.00909	0.19003	0.91050	0.00444	0.20709	0.89826
60%	25%	0.01693	0.30599	0.8550	0.00956	0.337	0.9014
70%	15%	0.00841	0.17893	0.92008	0.01047	0.28229	0.87301
80%	05%	0.00871	0.26119	0.86836	0.02631	0.59165	0.71077