Modeling Markov Switching ARMA-GARCH Neural Networks Models and an Application to Forecasting Stock Returns

Table 6

Markov switching ARMA-GARCH neural network models: out of sample results.

	MSE	RMSE
Model Group 1: MS-ARMA-GARCH-neural network models
MS-ARMA-GARCH	0.210599000	0.458911000 (6th)
MS-ARMA-GARCH-RNN	0.015436917	0.124245392 (2nd)
MS-ARMA-GARCH-RBF	0.023914916	0.154644483 (5th)
MS-ARMA-GARCH-ELMAN RNN	0.023667136	0.153841268 (4th)
MS-ARMA-GARCH-HYBRID MLP	0.015461623	0.124344774 (3rd)
MS-ARMA-GARCH-MLP	0.015333378	0.123828017 (1st)
Model Group 2: MS-ARMA-APGARCH-neural network models
MS-ARMA-APGARCH	0.17740000000000	0.42111000000000 (6th)
MS-ARMA-APGARCH-RNN	0.00000001628449	0.00012761067957 (4th)
MS-ARMA-APGARCH-RBF	0.00000002662825	0.00016318165387 (5th)
MS-ARMA-APGARCH-ELMAN RNN	0.00000001389858	0.00011789225071 (3rd)
MS-ARMA-APGARCH-HYBRID MLP	0.00000001331575	0.00011539391162 (1st)
MS-ARMA-APGARCH-MLP	0.00000001333791	0.00011548986313 (2nd)
Model Group 3: MS-ARMA-FIAPGARCH-neural network models
MS-ARMA-FIAPGARCH	0.17814000000000	0.42220660000000 (6th)
MS-ARMA-FIAPGARCH-RNN	0.00000001354729	0.00011639284333 (1st)
MS-ARMA-FIAPGARCH-RBF	0.00000002576468	0.00016051379977 (5th)
MS-ARMA-FIAPGARCH-ELMAN RNN	0.00000001521219	0.00012333770222 (4th)
MS-ARMA-FIAPGARCH-HYBRID MLP	0.00000001385396	0.00011770286713 (2nd)
MS-ARMA-FIAPGARCH-MLP	0.00000001389543	0.00011787886146 (3rd)