In this paper, we present a method for quantifying the degree of non-cooperation that exists among the target members of the neural network training set. Both the network architecture and the training algorithm are taken into consideration while computing non-cooperation measures. Based on these measures the network automatically partitions into several identical networks and each partition learns a subset of the targets. The partitioning takes place only when necessary and when needed the computation for partitioning is minimal. Each network is simple with only one hidden layer and currently has only one node in the output layer. A fusion network combines partial results to produce the final response. Simulation results indicate that the method is robust and capable of self organization to overcome the ill effects of non-cooperating targets in the training set, thereby reducing training time significantly.