BSc Hons (Melbourne); PhD (Melbourne), AStat
Professor Rob Hyndman is an internationally renowned researcher in forecasting and applied statistics. He has published three books and more than 120 papers in peer-reviewed international journals. His publications have been cited more 30000 times and his research and consulting work has had significant impact on society.
Professor Hyndman is known for his contribution to establishing forecasting as a multi-disciplinary science. He was the Director of the International Institute of Forecasters and Editor-in-Chief of the International Journal of Forecasting from 2005-2018. Under his editorship, the International Journal of Forecasting has become the premier research outlet in forecasting with a broad multi-disciplinary domain that includes science disciplines such as meteorology and power generation, and social science disciplines such as demography and macroeconomics.
Hyndman is well-known for his research in statistical forecasting methodology, especially in the areas of time series forecasting, forecast reconciliation, energy forecasting, and demographic forecasting.
In time series forecasting, Hyndman was one of the architects of the state space approach to exponential smoothing (along with Ralph Snyder, Keith Ord and Anne Koehler). This allowed likelihood-based model estimation, model selection, and prediction intervals to be computed. The resulting models are now known as ETS models, and are possibly the most widely used time series forecasting models in use today. He also developed with Yeasmin Khandakar an automatic time series forecasting algorithm based on ARIMA models which is very widely used in automatic forecasting implementations.
To measure the accuracy of time series forecasts across series on different scales, and with different characteristics, Hyndman developed the MASE (Mean Absolute Scaled Error). The MASE is now very widely used in time series forecast evaluations.
In 2011, Hyndman introduced (with coauthors) a new approach to reconciling very large numbers of forecasts from related time series using a least squares optimization method. Subsequent papers have extended the method in various ways and it is now widely used in forecasting practice.
Hyndman helped introduce functional data analysis into the demographic literature, and the functional data framework is now frequently used to analyse and forecast mortality, fertility, migration and population data.
In energy forecasting, Hyndman developed (with Shu Fan) a semi-parametric model for predicting extreme levels of demand several years in advance. This model has become a benchmark for subsequent empirical evaluations.
Making statistics useful for practical purposes has always been part of Professor Hyndman’s research agenda. To that end, he has been a strong proponent and developer of open source learning resources and open-source research tools. He has developed 50 open-source statistical software packages, which are all publicly available and are download millions of times every year. Jointly with George Athanasopoulos, he has developed an online textbook on forecasting that is being used in more than 200 countries.
Professor of Statistics, Monash University. 2003-
Moran Medal for Statistical Science, Australian Academy of Science, 2007
Elected Member, International Statistical Institute
Member, International Institute of Forecasters
Member, International Association for Statistical Computing
Member, Statistical Society of Australia
Member, International Society for Business and Industrial Statistics
1. Wickramasuriya, SL, G Athanasopoulos, and RJ Hyndman (2019). Optimal forecast reconciliation for hierarchical and grouped time series through trace minimization. J American Statistical Association 114(526), 804–819.
2. Forbes, J, D Cook, and RJ Hyndman (2020). Spatial modelling of the two-party preferred vote in Australian federal elections: 2001–2016. Australian & New Zealand Journal of Statistics 62(2), 168–185.
3. Makridakis, S, RJ Hyndman, and F Petropoulos (2020). Forecasting in social settings: the state of the art. International Journal of Forecasting 36(1), 15–28.
4. Talagala, PD, RJ Hyndman, and K Smith-Miles (2020). Anomaly detection in high-dimensional data. J Computational & Graphical Statistics. to appear.
5. Wang, E, D Cook, and RJ Hyndman (2020). Calendar-based graphics for visualizing people’s daily schedules. J Computational & Graphical Statistics 29 (3), 490–502.