Stock price series prediction optimization framework based on function information stacking and model averaging

Stock price series prediction has always been a hot issue in the field of quantitative finance. The commonly used models include ARIMA, GARCH, LSTM neural network and BP network. Aiming at these models, this paper proposes an optimization framework based on function information stacking model averaging. proposed method uses intra-day as auxiliary extracts functional features principal component analysis (PCA). Considering that underlying structure between characteristic variables residual obtained from original time is unknown, Stacking to enhance data reduce impact noise model. In addition, solve parameter problem model, averaging using distance covariance weighting deal with it. actual analysis, takes example explore effectiveness robustness method, results show certain competitiveness. Finally, can be improve other models.