A new method of stock index cross market forecasting based on deep learning and copula

Abstract

Aiming at the problem of stock index time series cross-market forecasting, a new method of stock index cross-market forecasting based on deep learning and Copula is constructed. From financial time series preprocessing, gradual decomposition, data dimensionality reduction, to deep learning feature extraction, a methodological paradigm for cross-market forecasting is proposed. It reveals the principles of data decomposition, risk estimation, deep learning integration to predict total risk, and risk ES estimation. The results show that the ES risk measurement model of the stock market based on deep learning is more accurate for risk capture and characterization, and has practical value. This has once again confirmed the great potential of deep learning in the financial field.