A hybrid approach of adaptive wavelet transform, long short-term memory and ARIMA-GARCH family models for the stock index prediction

作者：

Highlights：

• A new hybrid model which combines AWT, LSTM and ARIMA-GARCH-type models is proposed.

• A new hybrid model which combines AWT, LSTM and HAR(3) X-RV model is proposed.

• AWT-LSTM-ARMAX-FIEGARCH outperforms the benchmark models to predict the stock index.

• AWT-LSTM improves HAR(3) X-RV ability in prediction of the stock realized volatility.

摘要

•A new hybrid model which combines AWT, LSTM and ARIMA-GARCH-type models is proposed.•A new hybrid model which combines AWT, LSTM and HAR(3) X-RV model is proposed.•AWT-LSTM-ARMAX-FIEGARCH outperforms the benchmark models to predict the stock index.•AWT-LSTM improves HAR(3) X-RV ability in prediction of the stock realized volatility.

论文关键词：Stock index forecasting,Long short-term memory,Adaptive wavelet transform,FIGARCH and FIEGARCH models

论文评审过程：Received 25 November 2020, Revised 3 April 2021, Accepted 2 May 2021, Available online 5 May 2021, Version of Record 25 May 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115149