Empirical Comparison of Resampling Methods Using Genetic Neural Networks for a Regression Problem

In the paper the investigation of m-out-of-n bagging with and without replacement using genetic neural networks is presented. The study was conducted with a newly developed system in Matlab to generate and test hybrid and multiple models of computational intelligence using different resampling methods. All experiments were conducted with real-world data derived from a cadastral system and registry of real estate transactions. The performance of following methods was compared: classic bagging, out-of-bag, Efron’s .632 correction, and repeated holdout. The overall result of our investigation was as follows: the bagging ensembles created using genetic neural networks revealed prediction accuracy not worse than the experts’ method employed in reality.