Using Artificial Intelligence for Model Selection

We apply the optimization algorithm Adaptive Simulated Annealing (ASA) to the problem of analyzing data on a large population and selecting the best model to predict the probability that an individual with various traits will have a particular disease. We compare ASA with traditional forward and backward regression on computer simulated data. We find that the traditional methods of modeling are better for smaller data sets whereas a numerically stable ASA seems to perform better on larger and more complicated data sets.