Construction of Multi-step Price Forecasts in Commodity Markets Based on Qualitative and Quantitative Data Analysis Methods - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Construction of Multi-step Price Forecasts in Commodity Markets Based on Qualitative and Quantitative Data Analysis Methods

Résumé

The article proposes a method for constructing and correcting a multistep forecast for the year ahead (with a monthly breakdown) of prices for raw materials and products of industrial enterprises. The proposed approach consists in the formation of a price forecast taking into account 1) the price of the predicted indicator, the prices of goods participating in the product value chain, and macro indicators (time series); 2) information about the strength and direction of environmental factors affecting the market. Structured information about the effects of the external environment is the result of processing expert knowledge and hypotheses from heterogeneous information sources, through analysis and modeling on a cognitive map of the situation (CCS). We form a forecast by constructing an ensemble of time series models, each of which reflects the dependence of the target indicator on its past values and the prices of related products, the composition of which is determined by the results of cognitive modeling and time series analysis. Based on the results of monitoring on the cognitive map of the situation, conducting in order to analyze possible changes in the external environment and digital monitoring of prices, to identify changes in prices modes, we perform a forecast correction. The results obtained in this study show that the use of cognitive modeling and monitoring of changes improve the accuracy of predictions.
Fichier principal
Vignette du fichier
509923_1_En_68_Chapter.pdf (453.85 Ko) Télécharger le fichier
Origine : Accord explicite pour ce dépôt

Dates et versions

hal-04030405 , version 1 (16-03-2023)

Licence

Paternité

Identifiants

Citer

Zinaida K. Avdeeva, Elena A. Grebenyuk, Svetlana V. Kovriga. Construction of Multi-step Price Forecasts in Commodity Markets Based on Qualitative and Quantitative Data Analysis Methods. IFIP International Conference on Advances in Production Management Systems (APMS), Sep 2021, Nantes, France. pp.623-631, ⟨10.1007/978-3-030-85874-2_68⟩. ⟨hal-04030405⟩
10 Consultations
2 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More