IE-OPIM Joint Graduate Seminar
  • FENS
  • IE-OPIM Joint Graduate Seminar

You are here


IE-OPIM Joint Graduate Seminar: Esma Nur Çinicioğlu (İstanbul University)

How to Create Better Performing Bayesian Networks: A Heuristic Approach for Variable Selection

Date: Wednesday, April 22, 2015
Time: 13:40 – 14:30
Location: FENS L035


Abstract: The upsurge of popularity of Bayesian networks brings a parallel increase in research for structure learning algorithms of Bayesian networks from data sets. The ability of Bayesian networks to represent the probabilistic relationships between the variables is one of the main reasons of the rise in reputation of Bayesian networks as an inference tool. This also generates the major appeal of Bayesian networks for data mining. With the advancement and diversification of the structure learning algorithms, more variables may be incorporated to the learning process, bigger data sets may be used for learning, and inferences become faster even in the presence of continuous variables. On the other hand, though the quality of a learned network may be evaluated by many different aspects, the performance of the learned network very much depends on the selection of the variables to be included to the network. To assure the quality of the learned network structure, variable selection in Bayesian networks is necessary. Cinicioglu & Shenoy (2012) suggested an approach for variable selection in Bayesian networks where a score, Sj, is developed to assess each variable whether it should be included in the final Bayesian network. However, with this method the variables without parents or children are punished which affects the performance of the learned network. To eliminate that drawback, in this paper we develop a new score, NSj. We measure the performance of this new heuristic in terms of the prediction capacity of the learned network, its lift over marginal and evaluate its success by comparing it with the results obtained by the previously developed Sj score. For the illustration of the developed heuristic and comparison of the results credit score data is used.


Bio: Esma Nur Çinicioğlu is an Assistant Professor at Istanbul University, School of Business. Since December 2014  she is also acting as the MBA program coordinator of Istanbul University, Graduate School of Business. She received her bachelor degree from Marmara University in Management of Information Systems in 2002 and her PhD degree from University of Kansas  in Decision Sciences in 2008. She is the recipient of Max E. Fessler Dissertation Award for her dissertation titled "On Solving Stochastic PERT networks and Using RFIDs for Operations Management".  Her research interests include uncertainty management, belief functions theory, probabilistic graphical models with a special emphasis given to inference methods and variable selection heuristics in Bayesian networks. Her refereed articles have appeared in a variety of journals including Annals of Operations Research, International Journal of Approximate Reasoning, Expert Systems with Applications.