Two layers of neural network and it''s application in traditional Chinese medicine intelligent diagnosis
两层级神经网络及在中医智能诊断中的应用

By research on its characteristics, the traditional Chinese medicine ( TCM) clinical data are divided into two layer-grades: low layer-grade of data and high layer-grade of data. And every layer-grade data are divided into global data and local data. Based on this concept, built a two layer-grade neural network, which its low layer-grade sub-neural network calculated low layer-grade data and its high layer-grade sub-neural network comprehensively processed the high layer-grade data and the output of the low layer-grade sub-neural network. In this structure, not only was the TCM categorical identification problem effectively described, but also the calculation was simplified and the convergent learning speeded up. The result of experiments indicates that the two-layer neural network can be well applied to the TCM categorical identification intelligent calculation with complex data relationships.