Development and evaluation of probabilistic forecasting methods for small area populations

Planning and development decisions in both the government business sectors often require small area population forecasts. Unfortunately, current methods produce forecasts that are inaccurate, particularly for remote areas those with smaller populations. Such inaccuracy necessitates evaluation of to forecast communicate uncertainty, however, little research has been conducted this domain In paper, we evaluate a set probabilistic forecasting which include Autoregressive integrated moving average, Exponential Smoothing, THETA, LightGBM XGBOOST, point 80% prediction intervals Australian SA2 We also investigate combine ensemble Our results show individual generally underestimate uncertainty. Combining improves overall accuracy coverage their intervals, still tends be less than expected all but most conservative combination method.