05807906

a

05807906 Bernard, Simon Heutte, Laurent Adam, S\'ebastien A study of strength and correlation in random forests. Huang, De-Shuang (ed) et al., Advanced intelligent computing theories and applications. 6th international conference on intelligent computing, Changsha, China, August 18--21, 2010. Proceedings. Berlin: Springer (ISBN 978-3-642-14830-9/pbk; 978-3-642-14831-6/ebook). Communications in Computer and Information Science 93, 186-191 (2010). 2010 Berlin: Springer EN random forests ensemble of classifiers strength correlation

doi:10.1007/978-3-642-14831-6_25

Summary: In this paper we present a study on the Random Forest (RF) family of classification methods, and more particularly on two important properties called strength and correlation. These two properties have been introduced by Breiman in the calculation of an upper bound of the generalization error. We thus propose to experimentally study the actual relation between these properties and the error rate in order to confirm and extend the Breiman theoretical results. We show that the error rate statistically decreases with the joint maximization of the strength and minimization of the correlation, and this for different sizes of RF.