Skip to main content
Figure 1 | BioMedical Engineering OnLine

Figure 1

From: Statistical learning methods as a preprocessing step for survival analysis: evaluation of concept using lung cancer data

Figure 1

Modified probabilistic neural network stochastic training and z generation flow diagram. This schema shows the modified probabilistic neural network (PNN) training flow for the differential evolution (DE) sigma-weight vector construction, competition, and feedback from the g to the g+1 populations. The sigma-weight vectors x g and u g compete for the next generation. The receiver operating characteristic curve area (Az) from the stochastic cross-validation is derived with ensemble averaging to reduce the chance of passing outliers back to the vector competition. When g = G, the evolution stops and the sigma-weights are used in the PNN to generate z for each patient stochastically with ensemble averaging. The z quantities are then passed to the survival and logistic regression analyses.

Back to article page