Abstract
The sum of forecasts of disaggregated time series is often required to equal the forecast of the aggregate, giving a set of coherent forecasts. The least squares solution for finding coherent forecasts uses a reconciliation approach known as MinT, proposed by Wickramasuriya, Athanasopoulos, and Hyndman (2019). The MinT approach and its variants do not guarantee that the coherent forecasts are non-negative, even when all of the original forecasts are non-negative in nature. This has become a serious issue in applications that are inherently non-negative such as with sales data or tourism numbers. While overcoming this difficulty, we reconsider the least squares minimization problem with non-negativity constraints to ensure that the coherent forecasts are strictly non-negative. The constrained quadratic programming problem is solved using three algorithms. They are the block principal pivoting (BPV) algorithm, projected conjugate gradient (PCG) algorithm, and scaled gradient projection algorithm. A Monte Carlo simulation is performed to evaluate the computational performances of these algorithms as the number of time series increases. The results demonstrate that the BPV algorithm clearly outperforms the rest, and PCG is the second best. The superior performance of the BPV algorithm can be partially attributed to the alternative representation of the weight matrix in the MinT approach. An empirical investigation is carried out to assess the impact of imposing non-negativity constraints on forecast reconciliation over the unconstrained method. It is observed that slight gains in forecast accuracy have occurred at the most disaggregated level. At the aggregated level, slight losses are also observed. Although the gains or losses are negligible, the procedure plays an important role in decision and policy implementation processes.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.References
Bai, J.: Inferential theory for factor models of large dimensions. Econometrica 71(1), 135171 (2003)
Barzilai, J., Borwein, J.M.: Two-point step size gradient methods. IMA J. Numer. Anal. 8(1), 141–148 (1988)
Berry, M.W., Browne, M., Langville, A.N., Paul Pauca, V., Plemmons, R.J.: Algorithms and applications for approximate nonnegative matrix factorization. Comput. Stat. Data Anal. 52, 155–173 (2007)
Bertero, M., et al.: Scaled gradient projection methods for astronomical imaging. EAS Publ. Ser. 59, 325–356 (2013)
Birgin, E.G., Martinez, J.M., Raydan, M.: Inexact spectral projected gradient methods on convex sets. IMA J. Numer. Anal. 23(4), 539–559 (2003)
Bonettini, S., Zanella, R., Zanni, L.: A scaled gradient projection method for constrained image deblurring. Inverse Probl. 25(1), 23 (2009)
Chatfield, C.: Time-Series Forecasting, 1st edn. Chapman and Hall, New York (2000)
Chen, D., Plemmons, R.J.: Non-negativity constraints in numerical analysis. In: Bultheel, A., Cools, R. (eds.) The Birth of Numerical Analysis, pp. 109–140. World Scientific Publishing, New Jersey (2009)
Figueiredo, M.A.T., Nowak, R.D., Wright, S.J.: Gradient projection for sparse reconstruction: application to compressed sensing and other inverse problems. IEEE J. Sel. Top. Signal Process. 1(4), 586–597 (2007)
Hyndman, R.J.: forecast: forecasting functions for time series and linear models. R package version 8.5 (2019). https://pkg.robjhyndman/forecast. Accessed Feb 2019
Hyndman, R.J., Khandakar, Y.: Automatic time series forecasting: the forecast package for R. J. Stat. Softw. 27(3), 1–22 (2008)
Hyndman, R.J., Ahmed, R.A., Athanasopoulos, G., Shang, H.L.: Optimal combination forecasts for hierarchical time series. Comput. Stat. Data Anal. 55, 2579–2589 (2011). https://doi.org/10.1016/j.csda.2011.03.006
Hyndman, R.J., Lee, A.J., Wang, E.: Fast computation of reconciled forecasts for hierarchical and grouped time series. Comput. Stat. Data Anal. 97, 16–32 (2016)
Johnson, N.L., Kotz, S., Balakrishnan, N.: Continuous univariate distributions. In: Barnett, V. (ed.) Wiley Series in Probability and Mathematical Statistics, vol. 1, 2nd edn. Wiley, New York (1994)
Júdice, J.J., Pires, F.M.: Bard-type methods for the linear complementarity problem with symmetric positive definite matrices. IMA J. Manag. Math. 2(1), 5168 (1989)
Judice, J.J., Pires, F.M.: A block principal pivoting algorithm for large-scale strictly monotone linear complementarity problems. Comput. Oper. Res. 21(5), 587–596 (1994)
Karjalainen, E.J., Karjalainen, U.P.: Component reconstruction in the primary space of spectra and concentrations. Alternating regression and related direct methods. Anal. Chim. Acta 250, 169–179 (1991)
Kim, J., Park, H.: Fast nonnegative matrix factorization: an active-set-like method and comparisons. SIAM J. Sci. Comput. 33(6), 3261–3281 (2011)
Kostreva, M.M.: Block pivot methods for solving the complementarity problem. Linear Algebra Appl. 21, 207–215 (1978)
Lawson, C.L., Hanson, R.J.: Solving Least Squares Problems. Prentice-Hall, New Jersey (1974)
Microsoft Corporation and Steve Weston: doParallel: foreach parallel adaptor for the ‘parallel’ package. R package version 1.0.15 (2019a). https://CRAN.R-project.org/package=doParallel. Accessed Feb 2019
Microsoft Corporation and Steve Weston: foreach: provides foreach looping construct. R package version 1.4.7 (2019b). https://CRAN.R-project.org/package=foreach. Accessed Feb 2019
Murty, K.G.: Note on a Bard-type scheme for solving the complementarity problem. OPSEARCH 11(2–3), 123–130 (1974)
Nocedal, J., Wright, S.J.: In: Mikosch, T.V., Resnick, S.I., Robinson, S.M. (eds.) Numerical Optimization, 2nd edn. Springer, New York (2006)
Turlach, B.A., Weingessel, A.: quadprog: functions to solve quadratic programming problems. R package version 1.5-7 (2019). https://CRAN.R-project.org/package=quadprog. Accessed Feb 2019
Turlach, B.A., Wright, S.J.: Quadratic programming. Wiley Interdiscip. Rev. Comput. Stat. 7(2), 153–159 (2015)
Wickramasuriya, S.L., Athanasopoulos, G., Hyndman, R.J.: Optimal forecast reconciliation of hierarchical and grouped time series through trace minimization. J. Am. Stat. Assoc. 114(526), 804–819 (2019)
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Wickramasuriya, S.L., Turlach, B.A. & Hyndman, R.J. Optimal non-negative forecast reconciliation. Stat Comput 30, 1167–1182 (2020). https://doi.org/10.1007/s11222-020-09930-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11222-020-09930-0