Zero-Shot Forecasting Mortality Rates: A Global Study
Gabor Petnehazi, Laith Al Shaggah, Jozsef Gall, Bernadett Aradi
TL;DR
This global study evaluates zero-shot mortality rate forecasting using two foundation-model approaches (CHRONOS and TimesFM) against traditional and ML baselines across 5-, 10-, and 20-year horizons for 50 countries and 111 age groups. CHRONOS shows competitive short- to mid-term accuracy, especially when fine-tuned on mortality data, while TimesFM underperforms; Random Forest trained on mortality data often yields the best overall results. The findings indicate that zero-shot forecasting holds promise but requires careful model selection and domain adaptation, with domain-specific fine-tuning delivering meaningful gains for long-horizon forecasts. The work underscores the practical potential of zero-shot mortality forecasting while highlighting limitations and avenues for expanding model diversity and optimization.
Abstract
This study explores the potential of zero-shot time series forecasting, an innovative approach leveraging pre-trained foundation models, to forecast mortality rates without task-specific fine-tuning. We evaluate two state-of-the-art foundation models, TimesFM and CHRONOS, alongside traditional and machine learning-based methods across three forecasting horizons (5, 10, and 20 years) using data from 50 countries and 111 age groups. In our investigations, zero-shot models showed varying results: while CHRONOS delivered competitive shorter-term forecasts, outperforming traditional methods like ARIMA and the Lee-Carter model, TimesFM consistently underperformed. Fine-tuning CHRONOS on mortality data significantly improved long-term accuracy. A Random Forest model, trained on mortality data, achieved the best overall performance. These findings underscore the potential of zero-shot forecasting while highlighting the need for careful model selection and domain-specific adaptation.