Souhaib Ben Taieb, James W. Taylor, Rob J. Hyndman

Many applications require forecasts for a hierarchy comprising a set of time series along with aggregates of subsets of these series. Although forecasts can be produced independently for each series in the hierarchy, typically this does not lead to coherent forecasts — the property that forecasts add up appropriately across the hierarchy. State-of-the-art hierarchical forecasting methods usually reconcile these independently generated forecasts to satisfy the aggregation constraints. A fundamental limitation of prior research is that it has looked only at the problem of forecasting the mean of each time series. We consider the situation where probabilistic forecasts are needed for each series in the hierarchy. We define forecast coherency in this setting, and propose an algorithm to compute predictive distributions for each series in the hierarchy. Our algorithm has the advantage of synthesizing information from different levels in the hierarchy through a sparse forecast combination and a probabilistic hierarchical aggregation. We evaluate the accuracy of our forecasting algorithm on both simulated data and large-scale electricity smart meter data. The results show consistent performance gains compared to state-of-the art methods

Download working paper

  Tag: data science

0 posts
March 7th, 2017

Coherent Probabilistic Forecasts for Hierarchical Time Series

Souhaib Ben Taieb, James W. Taylor, Rob J. Hyndman Abstract: Many applications require forecasts for a hierarchy comprising a set […]

February 14th, 2017

The Australian Macro Database: An online resource for macroeconomic research in Australia

Timur Behlul, Anastasios Panagiotelis, George Athanasopoulos, Rob J Hyndman, Farshid Vahid Abstract A website that encourages and facilities the use […]

February 14th, 2017

Macroeconomic forecasting for Australia using a large number of predictors

Bin Jiang, George Athanasopoulos, Rob J Hyndman, Anastasios Panagiotelis, Farshid Vahid Abstract A popular approach to forecasting macroeconomic variables is […]

January 13th, 2017

Visualising forecasting algorithm performance using time series instance spaces

Yanfei Kang1, Rob J Hyndman2, Kate Smith-Miles3 School of Statistics, Renmin University of China. Department of Econometrics and Business Statistics, […]

May 6th, 2016

Automatic foRecasting using R

Talk given at the Melbourne Data Science Initiative, 6 May 2016.  

February 29th, 2016

On sampling methods for costly multi-objective black-box optimization

Ingrida Steponavičė, Mojdeh Shirazi-Manesh, Rob J. Hyndman, Kate Smith-Miles and Laura Villanova In Advances in Stochastic and Deterministic Global Optimization, […]

February 19th, 2016

Dynamic Algorithm Selection for Pareto Optimal Set Approximation

Ingrida Steponavičė, Rob J Hyndman, Kate Smith-Miles, Laura Villanova Journal of Global Optimization (2016), pp.1-20. Abstract: This paper presents a meta-algorithm […]

February 4th, 2016

Forecasting uncertainty in electricity smart meter data by boosting additive quantile regression

Souhaib BenTaieb, Raphael Huser, Rob J. Hyndman and Marc G. Genton IEEE Transactions on Smart Grid (2016), 7(5), 2448-2455. Abstract: […]

January 25th, 2016

Probabilistic Energy Forecasting: Global Energy Forecasting Competition 2014 and Beyond

Tao Hong, Pierre Pinson, Shu Fan, Hamidreza Zareipour, Alberto Troccoli and Rob J Hyndman International Journal of Forecasting (2016), 32(3), 896–913. […]

August 17th, 2015

Machine learning bootcamp

A talk on time series forecasting for the Monash University Machine Learning Bootcamp. Demo R code

June 23rd, 2015

MEFM: An R package for long-term probabilistic forecasting of electricity demand

International Symposium on Forecasting Riverside, California   I will describe and demonstrate a new open-source R package that implements the […]

June 19th, 2015

Probabilistic forecasting of peak electricity demand

Southern California Edison Rosemead, California   Electricity demand forecasting plays an important role in short-term load allocation and long-term planning […]

June 4th, 2015

Probabilistic time series forecasting with boosted additive models: an application to smart meter data

By Souhaib Ben Taieb, Raphael Huser, Rob J Hyndman and Marc G Genton