A simulated annealing approach to clinical stage classification problem

摘要

In order to determine an appropriate remedy, it is important to classify the patients with a particular disease into a few stages from the viewpoint of disease advance. The clinical stage classification problem is to obtain such classifications. We have proposed a method using genetic algorithms (GA) to obtain statistically optimal classifications in the sense of the Akaike information criterion (AIC). However, the results obtained by the method are not always useful for the practical purposes. In this paper, a method of clinical stage classification using simulated annealing (SA) is proposed, by which an optimal or suboptimal classification with the given number of stages can be obtained.