关键词: 金融时间序列, 集成预测模型, 卷积神经网络, 长短期记忆, 自回归移动平均

Abstract: Accurately forecasting the financial time series data plays an essential role in the operation and management of financial markets. A new ARMA-CNN-LSTM model for forecasting financial time series data is proposed based on neural networks and ensemble learning. The above model combines the Convolutional Neural Network (CNN), Long Short-term Memory (LSTM) network, and Autoregressive Moving Average (ARMA) model in an integrated framework. To achieve the hybrid modeling of linear and nonlinear features in financial time series data, the CNN-LSTM model is used to model the spatiotemporal features of data and the ARMA model is used to model the autocorrelation features of data. The experimental results show that compared with the benchmark individual model, the proposed model has excellent accuracy and robustness in forecasting financial time series data.

Key words: Financial Time Series, Ensemble Forecasting Models, Convolutional Neural Network (CNN), Long Short-term Memory (LSTM), Autoregressive Integrated Moving Average (ARMA)