Probabilistic neural network classification for microarraydata

We propose Probabilistic Neural Networks (PNN) to explore classifying microarray data patterns in gene expressions. The approach employs representative data that has patterns already identified to conduct training and testing of the classification capabilities of the PNN. Most supervised learning neural network models require multiple training cycles, whereas the PNN builds its model from just one training cycle. We hypothesize that the PNN is an ideal model to use for a quick analysis of a dataset and can be used as a tool to conduct a 'sanity check' on other classification models. Results show that a high-level classification rate can be achieved with this model with low time and model complexity. Comparison study with Bayesian Neural Network with structural learning has also been provided.