A combination-based forecasting method for the M4-competition

Jaganathan, Srihari; Prakash, Pks

doi:10.1016/j.ijforecast.2019.03.030

Cited by 25 publications

(17 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We consider these methods state-of-the art based on the results of the M4 competition [2]. The top entries in that competition achieved their accuracy through ensembling these methods [4,16,17] or by combining them with neural networks [3].…”

Section: Local Benchmark Methodsmentioning

confidence: 99%

Principles and Algorithms for Forecasting Groups of Time Series: Locality and Globality

Montero-Manso¹,

Hyndman²

2020

Preprint

View full text Add to dashboard Cite

Forecasting groups of time series is of increasing practical importance. Some examples are: forecasting the demand for multiple products offered by a retailer, server loads within a data center or the number of completed ride shares in zones within a city. The local approach to this problem considers each time series separately and fits a function or model to each series. The global approach considers all time series as the same regression task and fits a single function to all series. For groups of similar time series, global methods outperform the more established local methods. However, there is recent empirical evidence showing surprisingly good performance of global models on heterogeneous groups of time series. This suggests a more general applicability of global methods, with major implications in forecasting theory and practice, in the form of more accurate tools for automatic forecasting and new scenarios to study. However, the evidence has been of empirical nature and a more fundamental study is required. In this paper, we formalize the setting of forecasting a set of time series with local and global learning algorithms, leading to the following contributions:• We show that global methods are not more restrictive than local methods for time series forecasting, a result which does not apply to sets of regression problems in general. Global and local methods can produce the same forecasts without any assumptions about similarity of the series in the set. This result shows that global models can succeed in a wider range of problems than previously thought. • We derive basic generalization bounds for local and global algorithms. We find that the complexity of local methods grows with the size of the set while it remains constant for global methods. Therefore a global algorithm can afford to be quite complex and still benefit from better generalization error than local methods for large datasets. These bounds serve to clarify and support recent experimental results in the area of time series forecasting, and guide the design of new algorithms. For the specific class of limited-memory autoregressive models, this bound leads to the design of global models with much larger memory than what is effective for local methods. • The findings are supported by an extensive empirical study. We show that purposely naïve algorithms derived from these principles, such as global linear models fit by least squares, deep networks or even high order polynomials, result in superior accuracy in benchmark datasets. In particular, global linear models show an unreasonable effectiveness, providing competitive forecasting accuracy with far fewer parameters than the simplest of local methods. Empirical evidence points towards global models being able to automatically learn long memory patterns and related effects that are only available to local models if introduced manually. Keywords Time SeriesConsider the problem of having to forecast many time series as a group. We might need to forecast tourist arrivals at all our resorts for n...

show abstract

Section: Local Benchmark Methodsmentioning

confidence: 99%

Principles and Algorithms for Forecasting Groups of Time Series: Locality and Globality

Montero-Manso¹,

Hyndman²

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…As for linear combinations, uniform weights are usually considered in cases where the performance of individual models is unknown for a given time series (Jaganathan and Prakash, 2020). Gunter and O ¨nder (2016) suggested an association of weights with forecasting accuracy.…”

Section: Forecast Combinationsmentioning

confidence: 99%

Combination forecasting using multiple attribute decision making in tourism demand

2022

View full text Add to dashboard Cite

Purpose This study aims to address three important issues of combination forecasting in the tourism context: reducing the restrictions arising from requirements related to the statistical properties of the available data, assessing the weights of single models and considering nonlinear relationships among combinations of single-model forecasts. Design Methodology Approach A three-stage multiple-attribute decision-making (MADM)-based methodological framework was proposed. Single-model forecasts were generated by grey prediction models for the first stage. Vlsekriterijumska Optimizacija I Kompromisno Resenje was adopted to develop a weighting scheme in the second stage, and the Choquet integral was used to combine forecasts nonlinearly in the third stage. Findings The empirical results for inbound tourism in Taiwan showed that the proposed method can significantly improve accuracy to a greater extent than other combination methods. Along with scenario forecasting, a good forecasting practice can be further provided by estimating ex-ante forecasts post-COVID-19. Practical Implications The private and public sectors in economies with high tourism dependency can benefit from the proposed method by using the forecasts to help them formulate tourism strategies. Originality Value This study contributed to presenting a MADM-based framework that advances the development of a more accurate combination method for tourism forecasting.

show abstract

“…We note that the best combination is usually obtained when the best individual methods are used and when the combination has the most diverse approaches [2,3]. For example, Jaganathan and Prakash propose two forecasting combination approaches methods based on [33]: (a) historical evidence of well-evaluated methods inspired in Uniform weight distribution [34] and (b) weights optimization [33]. Jaganathan and Prakash took 24 forecasting methods from several sources to design a combination that surpassed most of the M4-Competition methods.…”

Section: Related Workmentioning

confidence: 99%

FCTA: A Forecasting Combined Methodology with a Threshold Accepting Approach

Frausto–Solís

Rodriguez-Moya

González-Barbosa

et al. 2022

Mathematical Problems in Engineering

View full text Add to dashboard Cite

The combination of forecasting methods is a widespread technique. The most common technique for ensembling several individual methods is scoring. These ensemble methods have been useful for designing hybrid forecasting time series methods in several areas. However, more precise applications are required in modern times, and hybridizations using several ranking approaches have emerged to solve this problem. The main difficulty of this technique is finding the most suitable methodology to combine forecasting methods. This work presents a new methodology named FCTA (forecasting combined method with threshold accepting) for ensembling several forecasting methods. This methodology uses a Threshold Accepting algorithm for weighting individual predictions. FCTA starts from an initial weighting and aims to find the best ponderation of the individual methods by optimizing the precision of the global prediction. For testing FCTA, we selected a dataset taken from M4-Makridakis-competition, and we compared it with the best individual forecasting methods. FCTA is also compared with other successful methodologies. The experimentation shows that FCTA surpasses the best M4 individual methods and is equivalent or better than the best methodologies of the area.

show abstract

A combination-based forecasting method for the M4-competition

Cited by 25 publications

References 20 publications

Principles and Algorithms for Forecasting Groups of Time Series: Locality and Globality

Principles and Algorithms for Forecasting Groups of Time Series: Locality and Globality

Combination forecasting using multiple attribute decision making in tourism demand

FCTA: A Forecasting Combined Methodology with a Threshold Accepting Approach

Contact Info

Product

Resources

About