A targeted Bayesian network learning for classification. Issue 3 (4th May 2019)

• Record Type: Journal Article
• Authors: Gruber, A.; Ben-Gal, I.
• Is Part Of: Quality technology & quantitative management. Volume 16:Issue 3(2019)
• Page Start: 243
• Page End: 261
• Publication Date: 2019-05-04
• Subjects: Bayesian classifiers -- complexity–accuracy trade-off -- information theory -- AI -- machine learning -- target-oriented learning; Quality control -- Periodicals; Quality control -- Statistical methods -- Periodicals; Industrial management -- Periodicals; Industrial management; Management -- Research -- Methodology -- Periodicals; Qualitative research -- Periodicals; Management; Quality control; Quality control -- Statistical methods; Periodicals; 658.00721
• DOI: 10.1080/16843703.2017.1395109
• Languages: English
• ISSNs: 1684-3703

Abstract

Abstract: A targeted Bayesian network learning (TBNL) method is proposed to account for a classification objective during the learning stage of the network model. The TBNL approximates the expected conditional probability distribution of the class variable. It effectively manages the trade-off between the classification accuracy and the model complexity by using a discriminative approach, constrained by information theory measurements. The proposed approach also provides a mechanism for maximizing the accuracy via a Pareto frontier over a complexity–accuracy plane, in cases of missing data in the data-sets. A comparative study over a set of classification problems shows the competitiveness of the TBNL mainly with respect to other graphical classifiers.