Structurally adaptive self-organizing neural trees
Tianli Li, L. Fang, A. Jennings · 2003
An architecture for an adaptive self-organizing neural tree is proposed. The adaptive neural tree adapts to the changing environment by adding and deleting nodes. It also performs parameter adaptation by constantly adjusting the connection weights. It has the successive approximation property which enables hierarchical classification and fast search implementation. An example is given to illustrate the adaptivity of the neural tree. The statistics of the learning behavior are also given.>