Probabilistic versus Regression Modeling for Disinfection Byproducts

Abstract
Probabilistic network approaches, including Bayesian networks (BN) and decomposable Markov networks (DMN) are graphical models in which a problem is structured as a set of parameters and probabilistic relationships between them. Probabilistic analyses have been effectively used to incorporate expert knowledge and historical data for revising the prior belief in the light of new evidence. In this chapter, the probabilistic approach is compared to traditional methods such as regression analysis for water quality predictions. The capabilities and advantages/disadvantages of DMN approach are described. A DMN through machine learning on the basis of historical data from the experiments is constructed. The results indicate that DMN is a better prediction model than multiple regression both theoretically and experimentally for these applications.
This paper is only available in PDF Format: View full text PDF