A neuro evolutionary algorithm for patient calibrated prediction of survival in Glioblastoma patients

•A neuro-evolutionary algorithm is proposed for predicting survival of GBM patients.•Modified genetic operators have been proposed for crossover and mutation.•An Autoencoder is used to compress the input data.•Taguchi method is used to tune parameters of the proposed neuroevolutionary algorithm.•The proposed approach has an excellent accuracy in predicting GBM patient survival.•The network can be used for survival prediction under different treatment schedules.