A self-organized network with a supervised training

This paper introduces a self-organized network with a supervisor for data clustering and recognition. The network consists of a self-organized module and a supervised module. The self-organized module has been mainly controlled by the information of input data. However, the proposed module is also trained by supervised learning. The supervised learning is applied to the supervised module first and then it is also applied to the self-organized module in order to realize higher generality and higher accuracy than conventional neural networks.