Моделиране на вероятностния избор при покупка
Modeling Probabilistic Choices When Buying

Summary/Abstract: This study systematizes the theoretical foundations of mathematical modeling of consumer behavior in the choice process in the context of random utility theory. The main attention is paid to some aspects of econometric estimation of discrete choice models. The main thesis is that based on the theory of random utility, you can create tools for predictive analytics of demand. The aim of the study is to develop a transparent and traceable analytical methodology for analyzing empirical data from stated preferences, providing opportunities for evaluating, analyzing and predicting consumer choice. To achieve this goal, a brief synopsis of the discrete choice theory was made, as well as a review of methods for evaluating and analyzing traditional econometric models with discrete dependent variables. Particular attention is drawn to the problems in specifying adequate binary and multinomial regression models, which are the core of analytical tools, and whose evaluation and verification is feasible with the help of open source software. Systematizing achievements in the field, recommendations are given for using predictive models for solving marketing problems. Methods for predictive analytics of consumer choice are proposed, including output of aggregated forecast estimates, building of market simulators for predictive modeling of market share, assessment of willingness-to-pay and price elasticity at the level of product alternatives.

• Details
• Contents