Machine Learning-Based Forecast of USD/CNH Exchange Rate Trends

This paper carries out, first, a preliminary prediction at the level of the change of foreign exchange rates, which is of fundamental importance for Macroeconomics. Exchange rate volatility will affect the core economic elements and then will have an effect on the cost of trade across borders as well as the profits for businesses. Using the dataset of 5 years' worth of RMB/USD exchange rate data from October 2020 to October 2025. The model includes 57 missing values. Both linear regression models and decision trees were employed for predicting data. A better fit predictive model is found through a comparison of performance. Test the result, which gets us a high coefficient of determination R2=0.7326, thus making the model able to explain 73.26% of the exchange rate's changes and with an exact predictable function. Current model projects the exchange rate will be 7.4167 and forecasts it will rise to 7.6197 one year later. This causes a 4.48% appreciation of the USD, and as shown, they can increase the accuracy of an exact forecast from a short-term forecast of the exchange rate.