Robust Exponential Smoothing of Multivariate Time Series

Croux, Christophe; Gelper, Sarah; Mahieu, Koen

doi:10.2139/ssrn.1512492

Cited by 7 publications

(6 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The difference between the real value and the forecasted value of a time series phenomenon is called the forecasting error (Croux et al , 2010). Regardless of which method is used to obtain the forecasting value of ( D t ) for single or multiple periods, the actual value will remain different from the forecasted value (Baykal–Gürsoy and Erkip, 2010).…”

Section: Theoretical Backgroundmentioning

confidence: 99%

Supply chain management and market responsiveness: a simulation study

Aljanabi

Ghafour

2020

JBIM

View full text Add to dashboard Cite

Purpose This study aims to provide a practical solution to the relationship between supply chain (SC) integration and market responsiveness (MR). A method is proposed to integrate SC and MR parameters, namely, product supply and demand in the context of low-value commodities (e.g. cement). Design/methodology/approach Simulation and forecasting approaches are adopted to develop a potential procedure for addressing demand during lead time. To establish inventory measurements (safety stock and reorder level) and increase MR and the satisfaction of customer’s needs, this study considers a downstream SC including manufacturers, depots and central distribution centers that satisfies an unbounded number of customers, which, in turn, transport the cement from the industrialist. Findings The demand during lead time is shown to follow a gamma distribution, a rare probability distribution that has not been considered in previous studies. Moreover, inventory measurements, such as the safety stock, depending on the safety factor under a certain service level (SL), which enables the SC to handle different responsiveness levels in accordance with customer requests. In addition, the quantities of the safety stock and reorder point represent an optimal value at each position to avoid over- or understocking. The role of SC characteristics in MR has largely been ignored in existing research. Originality/value This study applies SC flexibility analyzes to overcome the obstacles of analytical methods, especially when the production process involves probabilistic variables such as product availability and demand. The use of an efficient method for analyzing the forecasting results is an unprecedented idea that is proven efficacious in investigating non-dominated solutions. This approach provides near-optimal solutions to the trade-off between different levels of demand and the SC responsiveness (SLs) with minimal experimentation times.

show abstract

Section: Theoretical Backgroundmentioning

confidence: 99%

Supply chain management and market responsiveness: a simulation study

Aljanabi

Ghafour

2020

JBIM

View full text Add to dashboard Cite

show abstract

“…Loisel et al, 2008), time series modelling (e.g. Croux et al, 2010) and claims reserving (e.g. Verdonck and Debruyne, 2011).…”

Section: Robustness Issuesmentioning

confidence: 99%

Optimal risk transfer under quantile-based risk measurers

Asimit

Badescu

Verdonck

2013

Insurance: Mathematics and Economics

View full text Add to dashboard Cite

The classical problem of identifying the optimal risk transfer from one insurance company to multiple reinsurance companies is examined under some quantilebased risk measure criteria. We develop a new methodology via a two-stage optimisation procedure which allows us to not only recover some existing results in the literature, but also makes possible the analysis of high dimensional problems in which the insurance company diversifies its risk with multiple reinsurance counter-parties, where the insurer risk position and the premium charged by the reinsurers are functions of the underlying risk quantile. Closed form solutions are elaborated for some particular settings, although numerical methods for the second part of our procedure represent viable alternatives for the ease of implementing it in more complex scenarios. Furthermore, we discuss some approaches to obtain more robust results.

show abstract

“…An extension, which is suitable for trend models, is the double exponential smoother (ES):

\begin{matrix} left \\ {\overset{s}{true}}_{t} = λ {\overset{s}{true}}_{t - 1} + (1 - λ) Y_{t} \\ {\overset{g}{true}}_{t} = λ {\overset{g}{true}}_{t - 1} + (1 - λ) {\overset{s}{true}}_{t} \end{matrix}

where λ ∈ (0,1] is a weighting factor, which gives more weight to recent observations. The first equation can be applied to forecasting as

{\overset{true^}{Y}}_{t + 1} = {\overset{s}{true}}_{t}

, and a multivariate robust version has been discussed in Croux et al .Also, double ES is concerned with forecasting. As in the linear trend model, the forecast function is the sum of level and slope components.…”

Section: Representation and Estimationmentioning

confidence: 99%

Sequential smoothing for turning point detection with application to financial decisions

Grillenzoni

2012

Appl Stoch Models Bus & Ind

View full text Add to dashboard Cite

A fundamental problem in financial trading is the correct and timely identification of turning points in stock value series. This detection enables to perform profitable investment decisions, such as buying-at-low and selling-at-high. This paper evaluates the ability of sequential smoothing methods to detect turning points in financial time series. The novel idea is to select smoothing and alarm coefficients on the gain performance of the trading strategy. Application to real data shows that recursive smoothers outperform two-sided filters at the out-of-sample level

show abstract

Robust Exponential Smoothing of Multivariate Time Series

Cited by 7 publications

References 21 publications

Supply chain management and market responsiveness: a simulation study

Supply chain management and market responsiveness: a simulation study

Optimal risk transfer under quantile-based risk measurers

Sequential smoothing for turning point detection with application to financial decisions

Contact Info

Product

Resources

About