中医八纲辨证神经网络结构优化研究

Optimization for Neural Network Structure of TCM Eight Principle Syndromes

针对中医八纲辨证神经网络结构的缺点,在中医八纲辨证神经网络权值优化算法(MIGA-1)的基础上,研制出一种优化中医八纲辨证神经网络结构的免疫遗传算法——MIGA-2算法。实验结果表明,以MIGA-2算法作为结构优化机制的中医八纲辨证神经网络具有更好的自我适应能力,减少了对标准样本病例模范性要求。

Aiming at the shortcomings of the neural network structure of TCM eight principle syndromes, we optimize the neural network structure on the basis of the modified immune genetic algorithm-1 （ MIGA-1 ） and develop a kind of immune genetic algofithm-MIGA-2. This algorithm helps the neural network of TCM eight principle syndromes better adapt itself and diminish the requirement for standard samples.