Identification of the demand curve and forecasts in subsequent periods using the Metropolis-Hastings algorithm

Abstract

PURPOSE: The main purpose of the article is to identify the demand curve and to forecast demand in subsequent periods using the Metropolis-Hastings algorithm.

DESIGN/METHODOLOGY/APPROACH: The Metropolis-Hastings algorithm belonging to the Markov Chain Monte Carlo was used to identify the demand curve and to forecast the demand in subsequent periods. This method consists in generating (drawing) a sample in accordance with the modified distribution and the possibility of rejecting a new sample in case of insufficient improvement of the quality index.

FINDINGS: The results of the conducted research indicate that the presented solution of generating a sample in accordance with the modified distribution and the possibility of rejecting a new sample in the event of insufficient improvement of the quality index is effective in identifying and forecasting the demand.

PRACTICAL IMPLICATIONS: The algorithm presented in the article can be used to forecast stays taking into account the product life curve.

ORIGINALITY/VALUE: A novelty is the use of the Metropolis-Hastings algorithm to identify the demand curve and the forecast of demand in subsequent periods to determine the strategy of long-term products by analyzing the sales volume of the product.