Improved One-to-Many Record Linkage using One Class Clustering Tree

Record linkage is traditionally performed among the entities of same type. It can be done based on entities that may or may not share a common identifier. In this paper we propose a new linkage method that performs linkage between matching entities of different data types as well. The proposed technique is based on one-class clustering tree that characterizes the entities which are to be linked. The tree is built in such a way that it is easy to understand and can be transformed into association rules. The inner nodes of the tree consist of features of the first set of entities. The leaves of the tree represent features of the second set that are matching. The data is split using two splitting criteria. Also two pruning methods are used for creating one-class clustering tree. The proposed system results better in performance of precision and recall.