Enhanced One-Day-Ahead AUD/USD Exchange Rate Prediction using CatBoost Model

Abstract

The foreign exchange (Forex) market is highly liquid
and volatile, making accurate short-term forecasting both critical
and challenging. This study investigates one-day-ahead AUD/USD
exchange rate prediction using CatBoost, Random Forest (RF),
and Support Vector Machine (SVM) machine learning (ML)
models with continuous and discretized technical indicators. Ten
technical indicators were derived from 5,027 historical data points.
This is the first study to apply discrete technical indicators with
CatBoost for recent AUD/USD price forecasting. Results showed
that CatBoost achieved the highest accuracy (89.68%) and AUC
(0.9609) on the discretised dataset. Statistical test confirmed the
significance of CatBoost’s superior performance, highlighting its
potential to enhance predictive performance and support real-time
decision-making in Forex trading.