Investigation of random subspace and random forest regression models using data with injected noise

Abstract

The ensemble machine learning methods incorporating random subspace and random forest employing genetic fuzzy rule-based systems as base learning algorithms were developed in Matlab environment. The methods were applied to the real-world regression problem of predicting the prices of residential premises based on historical data of sales/purchase transactions. The accuracy of ensembles generated by the proposed methods was compared with bagging, repeated holdout, and repeated cross-validation models. The tests were made for four levels of noise injected into the benchmark datasets. The analysis of the results was performed using statistical methodology including nonparametric tests followed by post-hoc procedures designed especially for multiple NxN comparisons. © 2013 Springer-Verlag.