Details of Research Outputs

We develop Meta-Classifiers to detect financial frauds by combining several accurate and diverse stand-alone classifiers. Our results suggest that the Meta-Classifiers developed in our study can outperform the best stand-alone classifiers to detect fraudulent firms. We believe developing Meta-Classifiers can be a helpful technique to improve the predictive performance of models. Moreover, the methodology used to develop effective Meta-Classifiers in this study can also be replicated in other prediction related studies.