Ensemble Learning of Two Layers for Financial Time Series Prediction

Konferenz: ISCTT 2021 - 6th International Conference on Information Science, Computer Technology and Transportation
26.11.2021 - 28.11.2021 in Xishuangbanna, China

Tagungsband: ISCTT 2021

Seiten: 6Sprache: EnglischTyp: PDF

Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt

Hao, Chenchen (School of Electronics and Information Engineering, Beihang University, Beijing, China)
Gao, Qiang (School of Electronics and Information Engineering, Beihang University, Beijing, China & Hangzhou Innovation Institute, Beihang University, Hangzhou, China)

Various neural networks have been applied to financial time series prediction. Ensemble learning can take advantage of the different modeling capabilities of neural networks to achieve better prediction performance. In this paper, we propose a two-layer ensemble method for financial time series prediction. Firstly, Feedforward Neural Network (FNN), Long-Short Term Memory Neural Network (LSTM) and Convolutional Neural Network (CNN) are all trained on several training sets and multiple individual learners are obtained. The diversity between individual learners comes from different neural networks and different training sets. Secondly, individual learners of FNN, LSTM and CNN are combined separately to obtain three ensemble learners, which in turn are combined to obtain a two-layer ensemble learner. Experimental results demonstrate that the two-layer ensemble method can achieve better prediction accuracy and stability compared with the one-layer ensemble method in which the same individual learners are combined directly.