Learning Algorithm of Algebra Hyper Surface Neutral Network Model
This paper is about the research on the Learning Algorithm of Algebra Hyper Surface Neutral Network Model (AHSNNM), which is used to construct AHSNNM. AHSNNM is an extension of the simple perceptron model from the summing function. The summing function of AHSNNM is a polynomial in fact. The degree of the polynomial and the coefficient of each term can be obtained easily and rapidly by learning. And the learning algorithm of AHSNNM is a self-adaptive method which determines the most appropriate degree of polynomial by itself. AHSNNM can be used for classification and prediction through choosing different activation function and learning rule. Moreover, for classification the algorithm use a clever method that labels the classes of samples with binary numbers for solving multi-class problem and unifying two-class problem with multi-class problem. The experiment results show that the learning algorithm of AHSNNM is efficient and accurate, and thus AHSNNM can effectively support important decision-making.
neutral network perceptron algebraic hyper surface classification prediction self-adaptive multi-class
Zhenyan Liu Yong Wang Liping Chen
School of Software Beijing Institute of Technology Beijing, China Institute of Computing Technology School of Software Beijing Institute of Technology Beijing, China National Engineering Research Center for Information Technology in Agriculture Beijing, China
国际会议
太原
英文
449-453
2010-10-22（万方平台首次上网日期，不代表论文的发表时间）