Research and Application of a New Artificial Immune Algorithm Which Based on SOM Neural Network

A new immune algorithm which is based on the SOM neural network has been proposed in this paper. It takes the antigen as the neural network input pattern and the neural network connection weight as the antibody. It is made up of three processes: the recognition and matching of antibody to antigen; the clone proliferation and mutation of the antibody; the formation of antibody memory matrix. This paper analyzes the application of the immune algorithm in the silicon content prediction and results show that it has fascinating ability in silicone content prediction