Classification from one class of examples for relational domains
Tushar Khot, Sriraam Natarajan, JUDE W. SHAVLIK · 2014
One-class classification approaches have been proposed in the literature to learn classifiers from examples of only one class. But these approaches are not directly ap-plicable to relational domains due to their reliance on a feature vector or a distance measure. We propose a non-parametric relational one-class classification approach based on first-order trees. We learn a tree-based distance measure that iteratively introduces new relational fea-tures to differentiate relational examples. We update the distance measure so as to maximize the one-class clas-sification performance of our model. We also relate our model definition to existing work on probabilistic com-bination functions and density estimation. We experi-mentally show that our approach can discover relevant features and outperform three baseline approaches. 1