Forecasting small area populations with Long Short-Term Memory Networks

Abstract

Local and state governments depend on small area population forecasts to make important decisions concerning the development of local infrastructure and services, including schooling, transportation, healthcare, energy, telecommunications, and water supply. Despite their importance, current methods often produce highly inaccurate forecasts, especially at the small area scale. Recent years have witnessed promising developments in time series forecasting using Machine Learning across a wide range of social and economic variables. However, almost no work has been undertaken to investigate their potential application in demography, particularly for small area population forecasting. In this paper we describe the development of two Long-Short Term Memory network architectures for work with small area populations. We employ the Keras Tuner to select layer unit numbers, vary the window width of input data, and apply a double training and validation regime which supports work with short time series and prioritises later sequence values for forecasts. These methods are transferable and can be applied to other data sets. Retrospective small area population forecasts for Australia were created for the periods 2006-16 and 2011-16. Model performance was evaluated against actual data and two benchmark methods. The deep learning methods produced forecast errors comparable to the benchmark methods, markedly improving the 2006-based forecasts, but not 2011-based forecasts produced by one of the benchmark methods.

Dynamic Optimization of Investment Portfolio Based on Self-Attention Mechanism - LSTM Model Prediction

“…In recent years, the landscape of financial investment has undergone a transformative shift, embracing machine learning technologies to bolster the precision and efficacy of portfolio management strategies. Machine Learning has recently shown significant progress in forecasting time series across diverse socio-economic metrics [1]. A focal point of this transition is the deployment of deep learning models, notably the LSTM neural networks, tailored for financial statistics forecasting and consequent portfolio optimization.…”