Publication Date
2003
Abstract
This paper evaluates several modifications of the Simple Bayesian Classifier to enable estimation and inference over relational data. The resulting Relational Bayesian Classifiers are evaluated on three real-world datasets and compared to a baseline SBC using no relational information. The approach we call INDEPVAL achieves the best results. We use synthetic data sets to further explore performance as relational data characteristics vary.
Recommended Citation
Neville, Jennifer, "Simple Estimators for Relational Bayesian Classifiers" (2003). Computer Science Department Faculty Publication Series. 112.
Retrieved from https://scholarworks.umass.edu/cs_faculty_pubs/112
Comments
This paper was harvested from CiteSeer