Advantages of Combining Factorization Machine with Elman Neural Network for Volatility Forecasting of Stock Market

Table 3

Comparisons of prediction performances among different models for .

Evaluation errors	S&P 500			DJIA
	BPNN	Elman NN	FM-Elman	BPNN	Elman NN	FM-Elman
MAE	54.1822	46.8432	43.5088	321.3915	294.4472	280.9011
MAPE	5.1566	4.5231	4.1692	3.3196	3.0473	2.9379
RMSE	60.2121	53.8414	49.7034	382.7756	354.6576	348.8277
MAPE (100)	1.7813	1.5526	1.5273	1.1938	1.1733	1.1103
Evaluation errors	SSEC			SZI
	BPNN	Elman NN	FM-Elman	BPNN	Elman NN	FM-Elman
MAE	68.5672	53.6271	50.4433	287.2003	255.4868	228.2216
MAPE	2.4598	1.9369	1.8008	2.6363	2.2933	2.0552
RMSE	85.4527	72.682	69.7871	355.0629	324.7933	298.6005
MAPE (100)	2.8151	2.5933	2.5618	3.0594	2.8707	2.7645