Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: Integrative analysis and machine learning on cancer genomics data using the Cancer Systems Biology Database (CancerSysDB)

Fig. 2

Results of a cross validation of the random forest prediction of cancer types in the CancerSysDB. The predictions are based on a random forest learned on the training set comprising 6006 patients from 30 TCGA studies (Table 2). Displayed are the predictions of the classes in the 3085 patients in the training set. The accuracy strongly varies across the particular subclasses, but sums up to a total of 1521 correctly classified patients (49.3%)

Back to article page