single layer perception model that tries to replicate training data, trained with a stochastic update algorithm similar to implemented in genetic algorithms ; may overfit so recommended data size is 2-3 for rand data, time complexity may increase exponentially based on data set size
