tMissing data in large insurance datasets affects the learning and classification accuracies in predictivemodelling. Insurance datasets will continue to increase in size as more variables are added to aid inmanaging client risk and will therefore be