Classification performance of the three observed approaches.
收藏数据链接：
官方服务：
资源简介：
Four sets of boxplots represent predictive performance measured in Area under the ROC curve (AUC) for 1-Standard Error (1SE), boosted C5.0 decision trees (C5.0), glinternet (GLI) and a model using optimal lambda (OPT) setting obtained using cross-validation. Each set is obtained for a different setting of “Number of Discovered Interactions” (NDI)–i.e. 5, 10, 15 and 20 interactions.
创建时间：
2016-02-23
