From decision tree to heterogeneous decision forest : A novel chemometrics approach for structure-activity relationship modeling

摘要: The techniques of combining the predictions multiple classification models to produce a single model have been investigated for many years. In earlier applications, be combined developed by altering training set. use these so-called resampling techniques, however, enhance risk reducing predictivity and/or over fitting noise in data, which might result poorer prediction composite than individual models. this paper, we suggest novel approach, named Heterogenious Decision Forest (HDF), that combines Tree Each is using unique set descriptors. When similar predictive quality are HDF method, compared consistently and significantly improved both testing steps. An example will presented binding affinity 232 chemicals estrogen receptor.