A hybrid model for stock price prediction based on multi-view heterogeneous data

Financial Innovation

Table 4 The predicting accuracy of four models with different lags

Model	Introduction	Training (%)	Statistics	Lag0	Lag1
SVMMD	SVM based on market data	60	Average	–	0.5204
			Median	–	0.5147
		75	Average	–	0.5095
			Median	–	0.5050
		90	Average	–	0.5077
			Median	–	0.5000
SVMFN	SVM based on financial news	60	Average	0.5418	0.5291
			Median	0.5418	0.5240
		75	Average	0.5387	0.5328
			Median	0.5956	0.5275
		90	Average	0.6123	0.6170
			Median	0.6027	0.6180
SVMMV	SVM based on news and market data	60	Average	0.5533	0.5232
			Median	0.5411	0.5240
		75	Average	0.5545	0.5331
			Median	0.5464	0.5301
		90	Average	0.5536	0.5495
			Median	0.5556	0.5479
MVL-SVM	MVL-SVM based on news and market data	60	Average	0.8467	0.8504
			Median	0.8527	0.8562
		75	Average	0.8635	0.8588
			Median	0.8571	0.8634
		90	Average	0.8691	0.8741
			Median	0.8630	0.8767