Consumer Behaviour Timewise Dependencies Investigation by Means of Transition Graph

Abstract

The investigation of consumption behaviour on the level of every single person or some certain groups put forward some new tasks different from the behavioural analysis of the whole population. One of them is the problem of temporal peculiarities of consumer behaviour, for instance, how to find those, who react on some critical events faster than the others. It could be useful for identifying a focus-group which would show the tendency and help to make more accurate predictions for the rest of the population. A graph-based method of consumer’s behaviour analysis in the state space is developed in this research. The moments when the deviations from the usual behavioural trajectory occur are detected by incremental comparing the transition graph with its previous state. These moments collected for all customers help to separate the population by the delay time of their reaction to the critical situation. It’s also noticed that the velocity of the reaction is a personal feature of a customer, hence, this separation stays actual for different external events which cause the behavioural anomalies.