Machine learning advances for time series forecasting

Abstract

In this paper, we survey the most recent advances in supervised machine learning (ML) and high‐dimensional models for time‐series forecasting. We consider both linear and nonlinear alternatives. Among the linear methods, we pay special attention to penalized regressions and ensemble of models. The nonlinear methods considered in the paper include shallow and deep neural networks, in their feedforward and recurrent versions, and tree‐based methods, such as random forests and boosted trees. We also consider ensemble and hybrid models by combining ingredients from different alternatives. Tests for superior predictive ability are briefly reviewed. Finally, we discuss application of ML in economics and finance and provide an illustration with high‐frequency financial data.

Machine learning-based time series models for effective CO2 emission prediction in India

“…Let us know about some of them. In the first paper (Masini et al, 2021), the author survey the most recent advances in supervised machine learning (ML) and high dimensional models by considering both linear and non-linear methods for time-series forecasting and considering ensemble and hybrid models by combining ingredients from different alternatives. They also apply time series forecasting in the economic and financial fields.…”

Forecasting US Inflation Using Bayesian Nonparametric Models

“…Babb and Detmeister (2017) provide a useful review of the literature on nonlinear Phillips curves. A fast-growing literature evaluates the use of machine learning techniques for macroeconomic forecasting, with random forests (see Breiman (2001) and, e.g., Masini, Medeiros, and Mendes (2021), for a survey) performing particularly well, also during crisis times, in a variety of studies and for key variables such as GDP growth and inflation; see, e.g., Goulet , , Goulet Coulombe, Marcellino, and Stevanovic (2021), .…”

A machine learning approach to water quality forecasts and sensor network expansion: Case study in the Wabash River Basin, United States

“…Given the limitations detailed above, one promising approach to advance near‐term nutrient forecasting, particularly given the spatially sparse and temporally limited network of nutrient monitoring—is machine learning (ML). Forecasts based on ML approaches have been shown to perform well with sparse data sets while producing computationally efficient and accurate results (Lim & Zohren, 2021; Masini et al, 2020). ML models have the ability to predict complex non‐linear trends by learning from historical data.…”

“…ML models have the ability to predict complex non‐linear trends by learning from historical data. Indeed, applications of machine learning in time series forecasting are becoming increasingly common (e.g., Ahmed et al, 2010; Deb et al, 2017; Lim & Zohren, 2021; Masini et al, 2020; Papacharalampous et al, 2018; Sit et al, 2020; Voyant et al, 2017). In cases when sufficiently long or rich data sets are not available, such as the limited nutrient monitoring network in the Midwestern US, ML models may be trained on synthetic data and still gain operational benefit, a strategy that has shown promise in ecological forecasts (Hittmeir et al, 2019).…”