Fig. 14From: High dimensional model representation of log-likelihood ratio: binary classification with expression dataThe associated un-normalized risk scores, i.e., S(Xf) and \(S(X_{f_{i},f_{j}})\), of two test points. Genes and gene pairs which are not used in the decision rule automatically get a risk of zero. Note most genes/gene pairs assign a larger risk to the point in class 1 (high risk patients) to the point in class 0 (low risk patient)Back to article page