Comparative analysis of discretization methods in Bayesian networks. (January 2017)

• Record Type: Journal Article
• Authors: Nojavan A., Farnaz; Qian, Song S.; Stow, Craig A.
• Is Part Of: Environmental modelling & software. Volume 87(2017)
• Page Start: 64
• Page End: 71
• Publication Date: 2017-01
• Subjects: Bayesian networks -- Discretization -- Environmental modeling -- Equal interval -- Equal quantile -- Moment matching; Environmental monitoring -- Computer programs -- Periodicals; Ecology -- Computer simulation -- Periodicals; Digital computer simulation -- Periodicals; Computer software -- Periodicals; Environmental Monitoring -- Periodicals; Computer Simulation -- Periodicals; Environnement -- Surveillance -- Logiciels -- Périodiques; Écologie -- Simulation, Méthodes de -- Périodiques; Simulation par ordinateur -- Périodiques; Logiciels -- Périodiques; Computer software; Digital computer simulation; Ecology -- Computer simulation; Environmental monitoring -- Computer programs; Periodicals; Electronic journals; 363.70015118
• DOI: 10.1016/j.envsoft.2016.10.007
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: A key step in implementing Bayesian networks (BNs) is the discretization of continuous variables. There are several mathematical methods for constructing discrete distributions, the implications of which on the resulting model has not been discussed in literature. Discretization invariably results in loss of information, and both the discretization method and the number of intervals determines the level of such loss. We designed an experiment to evaluate the impact of commonly used discretization methods and number of intervals on the developed BNs. The conditional probability tables, model predictions, and management recommendations were compared and shown to be different among models. However, none of the models did uniformly well in all comparison criteria. As we cannot justify using one discretization method against others, we recommend caution when discretization is used, and a verification process that includes evaluating alternative methods to ensure that the conclusions are not an artifact of the discretization approach. Highlights: An experiment was designed to evaluate the impact of commonly used discretization methods and number of intervals on the developed Bayesian Networks (BNs). The conditional probability tables, and hence, model predictions and management recommendations differed among models. None of the models did uniformly well in all comparison criteria. It is suggested that discretization be used with caution or be avoided, when possible, and the BN models be modified to accommodate continuous variables. A verification process is recommended that includes evaluating alternative methods to ensure that the conclusions are not an artifact of the discretization approach.