Leveraging machine learning for Australian macroeconomic forecasting

作者：Boyuan Li, Zihui Deng, Gaurav Gupta, Yingxi Chen · 发表于：The Proceedings of ANZIAM · 年份：2026 · DOI：10.21914/anziamproc.v66.19641

Macroeconomic forecasting has traditionally relied on regression and time series methods, which have often struggled to outperform basic benchmarks. In this study, we explore the use of machine learning (ML) and deep learning (DL) techniques to forecast key macroeconomic indicators for Australia, specifically focusing on Gross Domestic Product growth, Consumer Price Index inflation, and the Interbank Overnight Cash Rate. Using a comprehensive dataset spanning from 1985 to 2023, we incorporate 16 predictors identified from previous literature. Our findings indicate that ML methods, particularly ensemble approaches such as Random Forest and XGBoost, deliver superior predictive performance compared to traditional time series models and \textsc{dl} methods. Although DL models such as Long Short-Term Memory (LSTM) and Convolutional Neural Network-LSTM were also tested, they did not achieve comparable accuracy in this context, highlighting the effectiveness of density-based ML approaches for macroeconomic forecasting. References F. R. Alharbi and D. Csala. A seasonal autoregressive integrated moving average with exogenous factors (SARIMAX) forecasting model-based time series approach. Intentions 7.4 (2022), p. 94. doi: 10.3390/inventions7040094 A. Alsharef, K. Aggarwal, Sonia, M. Kumar, and A. Mishra. Review of ML and AutoML solutions to forecast time-series data. Arc. Comput. Meth. Eng. 29.7 (2022), pp. 5297–5311. doi: 10.1007/s11831-022-09765-0 R. Ashley. On the usefulness of ...