Home > Research > Publications & Outputs > On the Issue of Choosing the Best Predictive Mo...

Links

Text available via DOI:

View graph of relations

On the Issue of Choosing the Best Predictive Model Based on Bayesian Principles

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNChapter

Published

Standard

On the Issue of Choosing the Best Predictive Model Based on Bayesian Principles. / Svetunkov, Sergey; Svetunkov, Ivan.
Understanding the Digital Transformation of Socio-Economic-Technological Systems . ed. / Tessalano Campos Devezas; Mohammed Ali Berawi; Sergey Evgenievich Barykin; Tatiana Kudryavtseva. Cham: Springer, 2024. p. 107-119 (Lecture Notes in Networks and Systems; Vol. 951).

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNChapter

Harvard

Svetunkov, S & Svetunkov, I 2024, On the Issue of Choosing the Best Predictive Model Based on Bayesian Principles. in T Campos Devezas, MA Berawi, SE Barykin & T Kudryavtseva (eds), Understanding the Digital Transformation of Socio-Economic-Technological Systems . Lecture Notes in Networks and Systems, vol. 951, Springer, Cham, pp. 107-119. https://doi.org/10.1007/978-3-031-56677-6_8

APA

Svetunkov, S., & Svetunkov, I. (2024). On the Issue of Choosing the Best Predictive Model Based on Bayesian Principles. In T. Campos Devezas, M. A. Berawi, S. E. Barykin, & T. Kudryavtseva (Eds.), Understanding the Digital Transformation of Socio-Economic-Technological Systems (pp. 107-119). (Lecture Notes in Networks and Systems; Vol. 951). Springer. https://doi.org/10.1007/978-3-031-56677-6_8

Vancouver

Svetunkov S, Svetunkov I. On the Issue of Choosing the Best Predictive Model Based on Bayesian Principles. In Campos Devezas T, Berawi MA, Barykin SE, Kudryavtseva T, editors, Understanding the Digital Transformation of Socio-Economic-Technological Systems . Cham: Springer. 2024. p. 107-119. (Lecture Notes in Networks and Systems). doi: 10.1007/978-3-031-56677-6_8

Author

Svetunkov, Sergey ; Svetunkov, Ivan. / On the Issue of Choosing the Best Predictive Model Based on Bayesian Principles. Understanding the Digital Transformation of Socio-Economic-Technological Systems . editor / Tessalano Campos Devezas ; Mohammed Ali Berawi ; Sergey Evgenievich Barykin ; Tatiana Kudryavtseva. Cham : Springer, 2024. pp. 107-119 (Lecture Notes in Networks and Systems).

Bibtex

@inbook{70395939c870415f9755f611e81c29e8,
title = "On the Issue of Choosing the Best Predictive Model Based on Bayesian Principles",
abstract = "One of the most important problems in business forecasting is the problem of variables selection. It arises in many contexts and in many applications and is typically solved either using cross-validation techniques or with information criteria. According to the latter approach, a researcher selects the predictive model that has the smallest information criterion value among the models considered in a pool. This approach showed its efficiency in many contexts, but it is oblivious of the performance of models several steps ahead and on the data the model has not seen (test set). We propose a new method for selecting the best predictive model, based on direct use of Bayes{\textquoteright} theorem. In this method, a priori and a posteriori probabilities of suitability of each model for forecasting are calculated. A coefficient is then calculated using geometric mean of these estimates. The model for which the proposed calculated coefficient has the maximum value is considered to be the most appropriate for the data and can be used for forecasting. A comparative analysis of the main methods for selecting the best forecast model carried out by the authors on the example of one hundred different time series using autoregressive models showed that on average the proposed method for selecting a model works better than the existing ones.",
keywords = "Autoregression, Forecasting, Information theory, Model building, Regression analysis",
author = "Sergey Svetunkov and Ivan Svetunkov",
year = "2024",
month = jun,
day = "21",
doi = "10.1007/978-3-031-56677-6_8",
language = "English",
isbn = "9783031566769",
series = "Lecture Notes in Networks and Systems",
publisher = "Springer",
pages = "107--119",
editor = "{Campos Devezas}, Tessalano and Berawi, {Mohammed Ali} and Barykin, {Sergey Evgenievich} and Tatiana Kudryavtseva",
booktitle = "Understanding the Digital Transformation of Socio-Economic-Technological Systems",

}

RIS

TY - CHAP

T1 - On the Issue of Choosing the Best Predictive Model Based on Bayesian Principles

AU - Svetunkov, Sergey

AU - Svetunkov, Ivan

PY - 2024/6/21

Y1 - 2024/6/21

N2 - One of the most important problems in business forecasting is the problem of variables selection. It arises in many contexts and in many applications and is typically solved either using cross-validation techniques or with information criteria. According to the latter approach, a researcher selects the predictive model that has the smallest information criterion value among the models considered in a pool. This approach showed its efficiency in many contexts, but it is oblivious of the performance of models several steps ahead and on the data the model has not seen (test set). We propose a new method for selecting the best predictive model, based on direct use of Bayes’ theorem. In this method, a priori and a posteriori probabilities of suitability of each model for forecasting are calculated. A coefficient is then calculated using geometric mean of these estimates. The model for which the proposed calculated coefficient has the maximum value is considered to be the most appropriate for the data and can be used for forecasting. A comparative analysis of the main methods for selecting the best forecast model carried out by the authors on the example of one hundred different time series using autoregressive models showed that on average the proposed method for selecting a model works better than the existing ones.

AB - One of the most important problems in business forecasting is the problem of variables selection. It arises in many contexts and in many applications and is typically solved either using cross-validation techniques or with information criteria. According to the latter approach, a researcher selects the predictive model that has the smallest information criterion value among the models considered in a pool. This approach showed its efficiency in many contexts, but it is oblivious of the performance of models several steps ahead and on the data the model has not seen (test set). We propose a new method for selecting the best predictive model, based on direct use of Bayes’ theorem. In this method, a priori and a posteriori probabilities of suitability of each model for forecasting are calculated. A coefficient is then calculated using geometric mean of these estimates. The model for which the proposed calculated coefficient has the maximum value is considered to be the most appropriate for the data and can be used for forecasting. A comparative analysis of the main methods for selecting the best forecast model carried out by the authors on the example of one hundred different time series using autoregressive models showed that on average the proposed method for selecting a model works better than the existing ones.

KW - Autoregression

KW - Forecasting

KW - Information theory

KW - Model building

KW - Regression analysis

U2 - 10.1007/978-3-031-56677-6_8

DO - 10.1007/978-3-031-56677-6_8

M3 - Chapter

SN - 9783031566769

T3 - Lecture Notes in Networks and Systems

SP - 107

EP - 119

BT - Understanding the Digital Transformation of Socio-Economic-Technological Systems

A2 - Campos Devezas, Tessalano

A2 - Berawi, Mohammed Ali

A2 - Barykin, Sergey Evgenievich

A2 - Kudryavtseva, Tatiana

PB - Springer

CY - Cham

ER -