ANALYSIS OF THE INFLUENCE ON CONSUMER BEHAVIOR BY APPLYING BAYESIAN NETWORK

Bayesian networks, known as belief networks, belong to the order of graphical probabilistic models. Th ese graphical structures are used in expressing knowled ge about the uncertain domain. Conditional probabil ities in the graphs are often estimated by using appropri ate statistical and computational methods. Hence, Bayesian networks combine the principles of graph t heory, probability theory, computer science and statistics. For Bayesian networks can be said that they enable an effective representation and also al low calculation of the joint probability distribution o ver a set of random variables. Application of Bayes ian networks in various fields such as economics, engin eering, medicine, etc., has become extremely popula r in the last decade. Graphical structures of these netw orks are very suitable for combining prior knowledg e and observed data. For this reason, the application of Bayesian network is also possible in the case of mi ssing data, when there is a need for knowledge of cause-e ffect relationship and for understanding of differe nt aspects of the problem with predicting future event s. Consumer behavior, as a topic that often opens u p new questions, also provides the ability to be perceive d in terms of probability theory with its simultane ous display across acyclic graph of Bayesian network. T hus, the Bayesian network is a good choice for researchers who are faced with complex problems in which one wants to come to conclusions that are not warranted logically but, rather, probabilistically.