Forecasting CNY exchange rate accurately is a challenging task due to its complex coupling nature, which includes market-level coupling from interactions with multiple financial markets, macro-level coupling from interactions with economic fundamentals and deep coupling from interactions of the two aforementioned kinds of couplings. This study develops a new deep coupled Long Short-Term Memory (LSTM) approach, namely DC-LSTM, to capture the complex couplings for USD/CNY exchange rate forecasting. In this approach, a deep structure consisting of stacked LSTMs is built to model the complex couplings. The experimental results with 10 years data indicate that the proposed approach significantly outperforms seven other benchmarks. The DC-LSTM is verified to be a useful tool to make wise investment decisions through a profitability discussion. The purpose in this paper is to clarify the importance of coupling learning for exchange rate forecasting, and the usefulness of deep coupled model to capture the couplings.

The carbon market is recognized as the most effective means for reducing global carbon dioxide emissions. Effective carbon price forecasting can help the carbon market to solve environmental problems at a lower economic cost. However, the existing studies focus on the carbon premium explanation from the perspective of return and volatility spillover under the framework of the mean-variance low-order moment. Specifically, the time-varying, high-order moment shock of market asymmetry and extreme policies on carbon price have been ignored. The innovation of this paper is constructing a new hybrid model, NAGARCHSK-GRU, that is consistent with the special characteristics of the carbon market. In the proposed model, the NAGARCHSK model is designed to extract the time-varying, high-order moment parameter characteristics of carbon price, and the multilayer GRU model is used to train the obtained time-varying parameter and improve the forecasting accuracy. The results conclude that the NAGARCHSK-GRU model has better accuracy and robustness for forecasting carbon price. Moreover, the long-term forecasting performance has been proved. This conclusion proves the rationality of incorporating the time-varying impact of asymmetric information and extreme factors into the forecasting model, and contributes to a powerful reference for investors to formulate investment strategies and assist a reduction in carbon emissions.