Model clustering by deterministic annealing.
Bart Bakker, Tom Heskes · Radboud Repository (Radboud University) · 1999
Although training an ensemble of neural network solutions increases the amount of information obtained from a system, large ensembles may be hard to analyze. Since data clustering is a good method to summarize large bodies of data, we will show in this paper how to use clustering on instances of neural networks. We will describe an algorithm based on deterministic annealing, which is able to cluster various types of data. As an example, we will apply the algorithm to instances of three different types of MLP's, trained to predict the time of death of ovarian cancer patients. 1 Introduction In neural network analysis a growing trend exists to not only train a network to find the best solution, but to create an ensemble of good solutions. For example, bootstrapping [1] leads to many different models by optimizing each model on a different sample of the training set, whereas the Bayesian approach creates a probability distribution of all solutions, from which may be sampled by Markov Cha...