An evolutionary approach to constructing prognostic models

A prognostic model is sought to determine whether or not patients suffering from an uncommon form of cancer will survive. Given a set of case historie s, we attempt to find the relative weightings of the different variables th at are used to describe the cases. Our first innovation is to use a diffusi on genetic algorithm (DGA) to find weightings which will give optimal survi val predictions. The DGA enables a number of criteria to be satisfied simul taneously, making it particularly suitable for model building. A further in novation is a method of representing synergies between interacting factors. The evolved model correctly predicts 90% of the survivors and 87% of death s, an improvement over the current model. More significantly, the method en ables a simple model to be evolved, one that produces well-balanced predict ions, and one that is relatively easy for clinicians to use. The method was validated by running it on a training set made up of 90% of the original d atabase and then studying the performance of the generated models on a test set consisting of the remaining 10% of the cases. (C) 1999 Elsevier Scienc e B.V. All rights reserved.