Iterative Classification in Relational Data

Relational data offer a unique opportunity for improving the c lassification accuracy o f statistical m odels. If two objects are related, inferring something about one object can aid inferences about the other. We present an iterative classification p rocedure that exploits this characteristic of relational data. This approach uses simple Bayesian classifiers in an iterative fashion, dynamically upd ating the attributes of some objects as inferences are made about related ob jects. Inferences made with h igh confidence in initial iterations are fed back into the data and are used to inform subsequent i nferences about related ob jects. We evaluate the performance of this approach on a binary classification task. Experiments indicate that it erative classification significantly increases accuracy when compared to a single-pass approach.