Fig. 1From: Improved high-dimensional prediction with Random Forests by the use of co-dataIllustration of the sources of data used in CoRF for the LNM example. First, a base RF is fitted on the training data. Its output, v ij , together with the co-data, is used to train the co-data model. From the co-data model, we obtain a probability per gene used for refitting on the training data. In an extra step we validate the results on GSE84846Back to article page