Decoding currency dynamics: A multiscale machine learning approach integrating economic indicators, ESG, and investor sentiment

  • 12 Views
  • 6 Downloads

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License

The foreign exchange market, characterized by high volatility and economic significance, requires accurate predictive models. This study investigates the application of the Temporal Fusion Transformer (TFT), enhanced with Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), for forecasting major foreign exchange (forex) currency pairs: USD/EUR, USD/JPY, USD/CNY, USD/AUD, and USD/INR. The proposed framework integrates a wide range of economic indicators, which include interest rate differentials, GDP growth, and trade balances, alongside investor sentiment derived from Twitter and ESG-related news sentiment. By addressing the non-linear, multiscale nature of forex time series, the CEEMDAN-TFT model facilitates improved signal decomposition and interpretability. Empirical results indicate that the model demonstrates competitive performance across all five currency pairs, with the USD/EUR pair exhibiting the highest predictive accuracy. Other pairs, exhibiting good predictive accuracy, include USD/JPY and USD/CNY, underscoring the model’s adaptability to varying economic contexts. Performance is assessed using multiple error metrics, and the model is benchmarked against standard neural network approaches (MLP, RNN, LSTM, CNN). Variable importance analysis highlights the dynamic influence of interest rates, investor sentiment, and ESG factors across different market regimes. This study provides empirical evidence that including ESG and investor sentiment can improve the accuracy of currency forecasting models. This study provides guidance and a framework for informed decision-making for traders, analysts, and policymakers.

view full abstract hide full abstract
    • Figure 1. Daily rate movements of the USD and EUR pair
    • Figure 2. Daily rate movements of the USD and JPY pair
    • Figure 3. Daily rate movements of the USD and CNY pair
    • Figure 4. Daily rate movements of the USD and AUD pair
    • Figure 5. Daily rate movements of the USD and INR pair
    • Figure 6. CEEMDAN decomposition of the USD/EUR exchange rate into Intrinsic Mode Functions (IMFs) and a residual trend
    • Figure 7. CEEMDAN decomposition of the USD/JPY exchange rate
    • Figure 8. CEEMDAN decomposition of the USD/CNY exchange rate
    • Figure 9. CEEMDAN decomposition of the USD/AUD exchange rate
    • Figure 10. CEEMDAN decomposition of the USD/INR exchange rate
    • Figure 11. Model performance across time horizons
    • Figure 12. Model forecast for USD/EUR
    • Figure 13. Model forecast for USD/JPY
    • Figure 14. Model forecast for USD/CNY
    • Figure 15. Model forecast for USD/AUD
    • Figure 16. Model forecast for USD/INR
    • Figure 17. Attention weights of key variables across currency pairs
    • Table 1. Variables employed in the forecasting model
    • Table 2. Descriptive statistics
    • Table 3. ADF, PP, and KPSS tests for a random walk
    • Table 4. CEEMDAN-TFT and other models in-sample results
    • Table 5. CEEMDAN-TFT and other model out-of-sample results
    • Conceptualization
      Sougata Banerjee
    • Data curation
      Sougata Banerjee
    • Formal Analysis
      Sougata Banerjee
    • Investigation
      Sougata Banerjee
    • Methodology
      Sougata Banerjee
    • Project administration
      Sougata Banerjee
    • Resources
      Sougata Banerjee
    • Software
      Sougata Banerjee
    • Supervision
      Sougata Banerjee
    • Validation
      Sougata Banerjee
    • Visualization
      Sougata Banerjee
    • Writing – original draft
      Sougata Banerjee
    • Writing – review & editing
      Sougata Banerjee