Forecasting Demand in the Commodity Market of Food Products Using Neural Networks


Nuru M. R.

15th International Conference on Application of Fuzzy Systems, Soft Computing and Artificial Intelligence Tools, ICAFS 2022, Budva, Montenegro, 26 - 27 August 2022, vol.610 LNNS, pp.122-129 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 610 LNNS
  • Doi Number: 10.1007/978-3-031-25252-5_20
  • City: Budva
  • Country: Montenegro
  • Page Numbers: pp.122-129
  • Keywords: Commodity market, Elasticity coefficient of demand, Forecasting demand, Neural networks, Radial basic neural network, Regression neural networks
  • Azerbaijan State University of Economics (UNEC) Affiliated: Yes

Abstract

When opening a new commercial enterprise, a correct predictive estimate of the demand for the goods of this firm should be given. This work is devoted to the use of neural networks for demand forecasting, by determining the demand elasticity coefficient, which will allow commercial enterprises to give a correct predictive estimate of the demand for their goods in the commodity market. It is known that forecasting is a probabilistic scientifically substantiated judgment about the trends of the phenomenon under study in the future. On the other hand, it should be noted that: - demand is an important factor in the study of the commodity market; - when opening a new commercial enterprise, a correct predictive assessment of the demand for the goods of a given company in the commodity market should be given; - demand and supply appear on the commodity market not spontaneously, but are formed and act according to the relevant laws; relevant laws; - human experience and intuition are among the main sources of information for forecasting demand and future market characteristics; - it is impossible to obtain statistical data in advance for a certain period of time. The paper considers the features of forecasting demand in the commodity market and substantiates the relevance of using neural networks (which is more adequate to the market model in the presence of uncertainty factors) for their forecasting than the methods of extrapolation of trends that are actively used in solving economic problems. The issue of choosing the type of neural network for approximating and predicting the elasticity coefficient of demand for food and non-food products is considered.