Model selection in reconciling hierarchical time series

Model selection has been proven an effective strategy for improving accuracy in time series forecasting applications. However, when dealing with hierarchical time series, apart from selecting the most appropriate forecasting model, forecasters have also to select a suitable method for reconciling the base forecasts produced for each series to make sure they are coherent. Although some hierarchical forecasting methods like minimum trace are strongly supported both theoretically and empirically for reconciling the base forecasts, there are still circumstances under which they might not produce the most accurate results, being outperformed by other methods. In this paper we propose an approach for dynamically selecting the most appropriate hierarchical forecasting method and succeeding better forecasting accuracy along with coherence. The approach, to be called conditional hierarchical forecasting, is based on Machine Learning classification methods and uses time series features as leading indicators for performing the selection for each hierarchy examined considering a variety of alternatives. Our results suggest that conditional hierarchical forecasting leads to significantly more accurate forecasts than standard approaches, especially at lower hierarchical levels.

[1]  Fotios Petropoulos,et al.  Exploring the sources of uncertainty: Why does bagging for time series forecasting work? , 2018, Eur. J. Oper. Res..

[2]  George Athanasopoulos,et al.  Hierarchical forecasts for Australian domestic tourism , 2009 .

[3]  Max A. Little,et al.  Highly comparative time-series analysis: the empirical structure of time series and their methods , 2013, Journal of The Royal Society Interface.

[4]  Evangelos Spiliotis,et al.  Are forecasting competitions data representative of the reality? , 2020 .

[5]  Richard Gerlach,et al.  Demand forecasting in supply chain: The impact of demand volatility in the presence of promotion , 2019, Comput. Ind. Eng..

[6]  Gene Fliedner,et al.  An investigation of aggregate variable time series forecast strategies with specific subaggregate time series statistical correlation , 1999, Comput. Oper. Res..

[7]  Xiaozhe Wang,et al.  Rule induction for forecasting method selection: Meta-learning the characteristics of univariate time series , 2009, Neurocomputing.

[8]  S. Kolassa.  Evaluating predictive count data distributions in retail sales forecasting , 2016 .

[9]  Rob J Hyndman,et al.  Machine learning applications in time series hierarchical forecasting , 2019, ArXiv.

[10]  Vijay Mahajan,et al.  New product forecasting models. Directions for research and implementation , 1988 .

[11]  Rob J. Hyndman,et al.  Fast computation of reconciled forecasts for hierarchical and grouped time series , 2016, Comput. Stat. Data Anal..

[12]  John E. Boylan,et al.  Judging the judges through accuracy-implication metrics: The case of inventory forecasting , 2010 .

[13]  Nick S. Jones,et al.  catch22: CAnonical Time-series CHaracteristics , 2019, Data Mining and Knowledge Discovery.

[14]  Fotios Petropoulos,et al.  An evaluation of simple versus complex selection rules for forecasting many time series , 2014 .

[15]  Leonard J. Tashman,et al.  Out-of-sample tests of forecasting accuracy: an analysis and review , 2000 .

[16]  Adel A. Ghobbar,et al.  Evaluation of forecasting methods for intermittent parts demand in the field of aviation: a predictive model , 2003, Comput. Oper. Res..

[17]  Guy Lapalme,et al.  A systematic analysis of performance measures for classification tasks , 2009, Inf. Process. Manag..

[18]  Svetlana Nikolicic,et al.  Forecasting hierarchical time series in supply chains: an empirical investigation , 2021, Int. J. Prod. Res..

[19]  Fotios Petropoulos,et al.  Cross-temporal aggregation: Improving the forecast accuracy of hierarchical electricity consumption , 2020, Applied Energy.

[20]  Jeffrey Sohl,et al.  Disaggregation methods to expedite product line forecasting , 1990 .

[21]  Jianru Xue,et al.  Increment Entropy as a Measure of Complexity for Time Series , 2015, Entropy.

[22]  Philip Hans Franses,et al.  The M3 competition: Statistical tests of the results , 2005 .

[23]  Ali Eshragh,et al.  Demand forecasting in the presence of systematic events: Cases in capturing sales promotions , 2019, International Journal of Production Economics.

[24]  Jerrold H. May,et al.  Determining an optimal hierarchical forecasting model based on the characteristics of the data set: Technical note , 2016 .

[25]  Gene Fliedner,et al.  Hierarchical forecasting: issues and use guidelines , 2001, Ind. Manag. Data Syst..

[26]  Fotios Petropoulos,et al.  Improving the forecasting performance of temporal hierarchies , 2019, PloS one.

[27]  Rob J. Hyndman,et al.  FFORMA: Feature-based forecast model averaging , 2020, International Journal of Forecasting.

[28]  Jan Fransoo,et al.  SKU demand forecasting in the presence of promotions , 2009, Expert Syst. Appl..

[29]  Tianqi Chen,et al.  XGBoost: A Scalable Tree Boosting System , 2016, KDD.

[30]  F. Petropoulos,et al.  On reconciling macro and micro energy transport forecasts for strategic decision making in the tanker industry , 2017 .

[31]  Xiaozhe Wang,et al.  Characteristic-Based Clustering for Time Series Data , 2006, Data Mining and Knowledge Discovery.

[32]  Bogdan Gabrys,et al.  Meta-learning for time series forecasting and forecast combination , 2010, Neurocomputing.

[33]  Fotios Petropoulos,et al.  forecast: Forecasting functions for time series and linear models , 2018 .

[34]  Robert Tibshirani,et al.  Classification by Pairwise Coupling , 1997, NIPS.

[35]  Nikolaos Kourentzes,et al.  Cross-temporal coherent forecasts for Australian tourism , 2019, Annals of Tourism Research.

[36]  Robert Fildes,et al.  Beyond forecasting competitions , 2001 .

[37]  Rajesh Piplani,et al.  Forecasting item-level demands: an analytical evaluation of top–down versus bottom–up forecasting in a production-planning framework , 2008 .

[38]  Fotios Petropoulos,et al.  'Horses for Courses' in demand forecasting , 2014, Eur. J. Oper. Res..

[39]  A. S. Dokuz,et al.  Wind power forecasting based on daily wind speed data using machine learning algorithms , 2019, Energy Conversion and Management.

[40]  Spyros Makridakis,et al.  The M5 competition: Background, organization, and implementation , 2021, International Journal of Forecasting.

[41]  Nigel Meade,et al.  Evidence for the selection of forecasting methods , 2000 .

[42]  Didrik Nielsen,et al.  Tree Boosting With XGBoost - Why Does XGBoost Win "Every" Machine Learning Competition? , 2016 .

[43]  Everette S. Gardner,et al.  Exponential smoothing: The state of the art , 1985 .

[44]  Slawek Smyl,et al.  A hybrid method of exponential smoothing and recurrent neural networks for time series forecasting , 2020, International Journal of Forecasting.

[45]  Fotios Petropoulos,et al.  Forecasting in social settings: The state of the art , 2020, International Journal of Forecasting.

[46]  Evangelos Spiliotis,et al.  The M4 Competition: 100,000 time series and 61 forecasting methods , 2020 .

[47]  Kate Smith-Miles,et al.  Visualising forecasting algorithm performance using time series instance spaces , 2017 .

[48]  Fred Collopy,et al.  Rule-Based Forecasting: Development and Validation of an Expert Systems Approach to Combining Time Series Extrapolations , 1992 .

[49]  Fred Collopy,et al.  Automatic Identification of Time Series Features for Rule-Based Forecasting , 2001 .

[50]  Kurt Hornik,et al.  kernlab - An S4 Package for Kernel Methods in R , 2004 .

[51]  K. Strimmer,et al.  Statistical Applications in Genetics and Molecular Biology A Shrinkage Approach to Large-Scale Covariance Matrix Estimation and Implications for Functional Genomics , 2011 .

[52]  Rob J. Hyndman,et al.  Optimal Forecast Reconciliation for Hierarchical and Grouped Time Series Through Trace Minimization , 2018, Journal of the American Statistical Association.

[53]  Veronica Piccialli,et al.  A machine learning approach for forecasting hierarchical time series , 2021, Expert Syst. Appl..

[54]  Rob J Hyndman,et al.  A state space framework for automatic forecasting using exponential smoothing methods , 2002 .

[55]  Rob J. Hyndman,et al.  Hierarchical Forecasting , 2019 .

[56]  Fred Collopy,et al.  An Application of Rule-Based Forecasting to a Situation Lacking Domain Knowledge , 2000 .

[57]  Lotfi Tadj,et al.  Exact and approximate solution for optimal inventory control of two-stock with reworking and forecasting of demand , 2019, Oper. Res..

[58]  Fotios Petropoulos,et al.  Hierarchical forecast reconciliation with machine learning , 2020, Appl. Soft Comput..

[59]  Sotirios Chatzis,et al.  Forecasting stock market crisis events using deep and statistical machine learning techniques , 2018, Expert Syst. Appl..

[60]  Amir F. Atiya,et al.  Why does forecast combination work so well? , 2020, International Journal of Forecasting.

[61]  Rob J. Hyndman,et al.  Optimal combination forecasts for hierarchical time series , 2011, Comput. Stat. Data Anal..

[62]  Byron J. Dangerfield,et al.  Top-down or bottom-up: Aggregate versus disaggregate extrapolations , 1992 .

[63]  Rajesh Piplani,et al.  On the effectiveness of top-down strategy for forecasting autoregressive demands , 2007 .

[64]  Fotios Petropoulos,et al.  Probabilistic forecast reconciliation with applications to wind power and electric load , 2019, Eur. J. Oper. Res..