Fast and accurate yearly time series forecasting with forecast combinations

It has long been known that combination forecasting strategies produce superior out-of-sample forecasting performances. In the M4 forecasting competition, a very simple forecast combination strategy achieved third place on yearly time series. An analysis of the ensemble model and its component models suggests that the competitive accuracy comes from avoiding poor forecasts, rather than from beating the best individual models. Moreover, the simple ensemble model can be fitted very quickly, can easily scale horizontally with additional CPU cores or a cluster of computers, and can be implemented by users very quickly and easily. This approach might be of particular interest to users who need accurate yearly forecasts without being able to spend significant time, resources, or expertise on tuning models. Users of the R statistical programming language can access this modeling approach using the “forecastHybrid” package.     

A combination-based forecasting method for the M4-competition

Several researchers (Armstrong, 2001; Clemen, 1989; Makridakis and Winkler, 1983) have shown empirically that combination-based forecasting methods are very effective in real world settings. This paper discusses a combination-based forecasting approach that was used successfully in the M4 competition. The proposed approach was evaluated on a set of 100K time series across multiple domain areas with varied frequencies. The point forecasts submitted finished fourth based on the overall weighted average (OWA) error measure and second based on the symmetric mean absolute percent error (sMAPE).     

Forecasting monthly and quarterly time series using STL decomposition

This paper is a re-examination of the benefits and limitations of decomposition and combination techniques in the area of forecasting, and also a contribution to the field, offering a new forecasting method. The new method is based on the disaggregation of time series components through the STL decomposition procedure, the extrapolation of linear combinations of the disaggregated sub-series, and the reaggregation of the extrapolations to obtain estimates for the global series. Applying the forecasting method to data from the NN3 and M1 Competition series, the results suggest that it can perform well relative to four other standard statistical techniques from the literature, namely the ARIMA, Theta, Holt-Winters’ and Holt’s Damped Trend methods. The relative advantages of the new method are then investigated further relative to a simple combination of the four statistical methods and a Classical Decomposition forecasting method. The strength of the method lies in its ability to predict long lead times with relatively high levels of accuracy, and to perform consistently well for a wide range of time series, irrespective of the characteristics, underlying structure and level of noise of the data.     

Forecasting the NN5 time series with hybrid models

We propose a simple way of predicting time series with recurring seasonal periods. Missing values of the time series are estimated and interpolated in a preprocessing step. We combine several forecasting methods by taking the weighted mean of forecasts that were generated with time-domain models which were validated on left-out parts of the time series. The hybrid model is a combination of a neural network ensemble, an ensemble of nearest trajectory models and a model for the 7-day cycle. We apply this approach to the NN5 time series competition data set.     

A refined parametric model for short term load forecasting

We present a refined parametric model for forecasting electricity demand which performed particularly well in the recent Global Energy Forecasting Competition (GEFCom 2012). We begin by motivating and presenting a simple parametric model, treating the electricity demand as a function of the temperature and day of the data. We then set out a series of refinements of the model, explaining the rationale for each, and using the competition scores to demonstrate that each successive refinement step increases the accuracy of the model’s predictions. These refinements include combining models from multiple weather stations, removing outliers from the historical data, and special treatments of public holidays.     

MLP ensembles improve long term prediction accuracy over single networks

This work describes an award winning approach for solving the NN3 Forecasting Competition problem, focusing on the sound experimental validation of its main innovative feature. The NN3 forecasting task consisted of predicting 18 future values of 111 short monthly time series. The main feature of the approach was the use of the median for combining the forecasts of an ensemble of 15 MLPs to predict each time series. Experimental comparison to a single MLP shows that the ensemble increases the performance accuracy for multiple-step ahead forecasting. This system performed well on the withheld data, having finished as the second best solution of the competition with an SMAPE of 16.17%.     

The M5 uncertainty competition: Results,findings and conclusions

This paper describes the M5 “Uncertainty” competition, the second of two parallel challenges of the latest M competition, aiming to advance the theory and practice of forecasting. The particular objective of the M5 “Uncertainty” competition was to accurately forecast the uncertainty distributions of the realized values of 42,840 time series that represent the hierarchical unit sales of the largest retail company in the world by revenue, Walmart. To do so, the competition required the prediction of nine different quantiles (0.005, 0.025, 0.165, 0.250, 0.500, 0.750, 0.835, 0.975, and 0.995), that can sufficiently describe the complete distributions of future sales. The paper provides details on the implementation and execution of the M5 “Uncertainty” competition, presents its results and the top-performing methods, and summarizes its major findings and conclusions. Finally, it discusses the implications of its findings and suggests directions for future research.     

Combining predictive distributions for the statistical post-processing of ensemble forecasts

Statistical post-processing techniques are now used widely for correcting systematic biases and errors in the calibration of ensemble forecasts obtained from multiple runs of numerical weather prediction models. A standard approach is the ensemble model output statistics (EMOS) method, which results in a predictive distribution that is given by a single parametric law, with parameters that depend on the ensemble members. This article assesses the merits of combining multiple EMOS models based on different parametric families. In four case studies with wind speed and precipitation forecasts from two ensemble prediction systems, we investigate the performances of state of the art forecast combination methods and propose a computationally efficient approach for determining linear pool combination weights. We study the performance of forecast combination compared to that of the theoretically superior but cumbersome estimation of a full mixture model, and assess which degree of flexibility of the forecast combination approach yields the best practical results for post-processing applications.     

Online hierarchical forecasting for power consumption data

This paper proposes a three-step approach to forecasting time series of electricity consumption at different levels of household aggregation. These series are linked by hierarchical constraints—global consumption is the sum of regional consumption, for example. First, benchmark forecasts are generated for all series using generalized additive models. Second, for each series, the aggregation algorithm ML-Poly, introduced by Gaillard, Stoltz, and van Erven in 2014, finds an optimal linear combination of the benchmarks. Finally, the forecasts are projected onto a coherent subspace to ensure that the final forecasts satisfy the hierarchical constraints. By minimizing a regret criterion, we show that the aggregation and projection steps improve the root mean square error of the forecasts. Our approach is tested on household electricity consumption data; experimental results suggest that successive aggregation and projection steps improve the benchmark forecasts at different levels of household aggregation.     

Exploring the representativeness of the M5 competition data

The main objective of the M5 competition, which focused on forecasting the hierarchical unit sales of Walmart, was to evaluate the accuracy and uncertainty of forecasting methods in the field to identify best practices and highlight their practical implications. However, can the findings of the M5 competition be generalized and exploited by retail firms to better support their decisions and operation? This depends on the extent to which M5 data is sufficiently similar to unit sales data of retailers operating in different regions selling different product types and considering different marketing strategies. To answer this question, we analyze the characteristics of the M5 time series and compare them with those of two grocery retailers, namely Corporación Favorita and a major Greek supermarket chain, using feature spaces. Our results suggest only minor discrepancies between the examined data sets, supporting the representativeness of the M5 data.     

The M5 competition: Background,organization, and implementation

The M5 competition follows the previous four M competitions, whose purpose is to learn from empirical evidence how to improve forecasting performance and advance the theory and practice of forecasting. M5 focused on a retail sales forecasting application with the objective to produce the most accurate point forecasts for 42,840 time series that represent the hierarchical unit sales of the largest retail company in the world, Walmart, as well as to provide the most accurate estimates of the uncertainty of these forecasts. Hence, the competition consisted of two parallel challenges, namely the Accuracy and Uncertainty forecasting competitions. M5 extended the results of the previous M competitions by: (a) significantly expanding the number of participating methods, especially those in the category of machine learning; (b) evaluating the performance of the uncertainty distribution along with point forecast accuracy; (c) including exogenous/explanatory variables in addition to the time series data; (d) using grouped, correlated time series; and (e) focusing on series that display intermittency. This paper describes the background, organization, and implementations of the competition, and it presents the data used and their characteristics. Consequently, it serves as introductory material to the results of the two forecasting challenges to facilitate their understanding.     

M5 accuracy competition: Results,findings, and conclusions

In this study, we present the results of the M5 “Accuracy” competition, which was the first of two parallel challenges in the latest M competition with the aim of advancing the theory and practice of forecasting. The main objective in the M5 “Accuracy” competition was to accurately predict 42,840 time series representing the hierarchical unit sales for the largest retail company in the world by revenue, Walmart. The competition required the submission of 30,490 point forecasts for the lowest cross-sectional aggregation level of the data, which could then be summed up accordingly to estimate forecasts for the remaining upward levels. We provide details of the implementation of the M5 “Accuracy” challenge, as well as the results and best performing methods, and summarize the major findings and conclusions. Finally, we discuss the implications of these findings and suggest directions for future research.     

GEFCom2012: Electric load forecasting and backcasting with semi-parametric models

We sum up the methodology of the team tololo for the Global Energy Forecasting Competition 2012: Load Forecasting. Our strategy consisted of a temporal multi-scale model that combines three components. The first component was a long term trend estimated by means of non-parametric smoothing. The second was a medium term component describing the sensitivity of the electricity demand to the temperature at each time step. We use a generalized additive model to fit this component, using calendar information as well. Finally, a short term component models local behaviours. As the factors that drive this component are unknown, we use a random forest model to estimate it.     

Interpreting the skill score form of forecast performance metrics

Performance measures of point forecasts are expressed commonly as skill scores, in which the performance gain from using one forecasting system over another is expressed as a proportion of the gain achieved by forecasting that outcome perfectly. Increasingly, it is common to express scores of probabilistic forecasts in this form; however, this paper presents three criticisms of this approach. Firstly, initial condition uncertainty (which is outside the forecaster’s control) limits the capacity to improve a probabilistic forecast, and thus a ‘perfect’ score is often unattainable. Secondly, the skill score forms of the ignorance and Brier scores are biased. Finally, it is argued that the skill score form of scoring rules destroys the useful interpretation in terms of the relative skill levels of two forecasting systems. Indeed, it is often misleading, and useful information is lost when the skill score form is used in place of the original score.     

Are forecasting competitions data representative of the reality?

Forecasters typically evaluate the performances of new forecasting methods by exploiting data from past forecasting competitions. Over the years, numerous studies have based their conclusions on such datasets, with mis-performing methods being unlikely to receive any further attention. However, it has been reported that these datasets might not be indicative, as they display many limitations. Since forecasting research is driven somewhat by data from forecasting competitions, it becomes vital to determine whether they are indeed representative of the reality or whether forecasters tend to over-fit their methods on a random sample of series. This paper uses the data from M4 as proportionate to the real world and compares its properties with those of past datasets commonly used in the literature as benchmarks in order to provide evidence on that question. The results show that many popular benchmarks of the past may indeed deviate from reality, and ways forward are discussed in response.     

An ensemble approach to GEFCom2017 probabilistic load forecasting

We present an ensembling approach to medium-term probabilistic load forecasting which ranked second out of 73 competitors in the defined data track of the GEFCom2017 qualifying match. In addition to being accurate, the ensemble method is highly scalable, due to the fact that it had to be applied to nine quantiles in ten zones and for six rounds. Candidate forecasts were generated using random settings for input data, covariates, and learning algorithms. The best candidate forecasts were averaged to create the final forecast, with the number of candidate forecasts being chosen based on their accuracy in similar validation periods.     

The M4 Competition: 100,000 time series and 61 forecasting methods

The M4 Competition follows on from the three previous M competitions, the purpose of which was to learn from empirical evidence both how to improve the forecasting accuracy and how such learning could be used to advance the theory and practice of forecasting. The aim of M4 was to replicate and extend the three previous competitions by: (a) significantly increasing the number of series, (b) expanding the number of forecasting methods, and (c) including prediction intervals in the evaluation process as well as point forecasts. This paper covers all aspects of M4 in detail, including its organization and running, the presentation of its results, the top-performing methods overall and by categories, its major findings and their implications, and the computational requirements of the various methods. Finally, it summarizes its main conclusions and states the expectation that its series will become a testing ground for the evaluation of new methods and the improvement of the practice of forecasting, while also suggesting some ways forward for the field.     

Forecasting with gradient boosted trees: augmentation,tuning, and cross-validation strategies: Winning solution to the M5 Uncertainty competition

Deep neural networks and gradient boosted tree models have swept across the field of machine learning over the past decade, producing across-the-board advances in performance. The ability of these methods to capture feature interactions and nonlinearities makes them exceptionally powerful and, at the same time, prone to overfitting, leakage, and a lack of generalization in domains with target non-stationarity and collinearity, such as time-series forecasting. We offer guidance to address these difficulties and provide a framework that maximizes the chances of predictions that generalize well and deliver state-of-the-art performance. The techniques we offer for cross-validation, augmentation, and parameter tuning have been used to win several major time-series forecasting competitions—including the M5 Forecasting Uncertainty competition and the Kaggle COVID19 Forecasting series—and, with the proper theoretical grounding, constitute the current best practices in time-series forecasting.     

Correlated daily time series and forecasting in the M4 competition

We participated in the M4 competition for time series forecasting and here describe our methods for forecasting daily time series. We used an ensemble of five statistical forecasting methods and a method that we refer to as the correlator. Our retrospective analysis using the ground truth values published by the M4 organisers after the competition demonstrates that the correlator was responsible for most of our gains over the naïve constant forecasting method. We identify data leakage as one reason for its success, due partly to test data selected from different time intervals, and partly to quality issues with the original time series. We suggest that future forecasting competitions should provide actual dates for the time series so that some of these leakages could be avoided by participants.     

Embrace the differences: Revisiting the PollyVote method of combining forecasts for U.S. presidential elections (2004 to 2020)

While combining forecasts is well-known to reduce error, the question of how to best combine forecasts remains. Prior research suggests that combining is most beneficial when relying on diverse forecasts that incorporate different information. Here, I provide evidence in support of this hypothesis by analyzing data from the PollyVote project, which has published combined forecasts of the popular vote in U.S. presidential elections since 2004. Prior to the 2020 election, the PollyVote revised its original method of combining forecasts by, first, restructuring individual forecasts based on their underlying information and, second, adding naïve forecasts as a new component method. On average across the last 100 days prior to the five elections from 2004 to 2020, the revised PollyVote reduced the error of the original specification by eight percent and, with a mean absolute error (MAE) of 0.8 percentage points, was more accurate than any of its component forecasts. The results suggest that, when deciding about which forecasts to include in the combination, forecasters should be more concerned about the component forecasts’ diversity than their historical accuracy.