| 摘要: |
| 提出一种用于多项式理想计算的理想同余神经元,其工作方式既不同于过去感知器输入加阈值的激活方式,也不同于通常意义下激活函数选取,且保持神经元的运算特性.以Grobner基计算为例,给出利用该神经元计算Grobner基神经网络描述性学习算法. |
| 关键词: 理想同余神经元 多项式理论 Grobner基 描述性学习算法 |
| DOI： |
| 投稿时间：2000-09-28修订日期：2000-12-28 |
| 基金项目: |
|
| Neural Network Approach to Polynomial Ideal Computation |
|
Zhou Yongquan
|
| (Dept. of Math. & Comp. Sci., Guangxi Univ. for Nationalities, Xixiangtang, Nanning, Guangxi, 530006, China) |
| Abstract: |
| The concept of polynomial ideal congruent neuron are proposed. It is different from both general perception and chosen stimulate function in working way, and is still characterized massively parallel architecture of the polynomial ideal computation. The polynomial ideal congruent neural network learning algorithm is discussed with an application of this neuron to computation of Grobner base. |
| Key words: indeal congruent neuron polynomial theory Grobner base polynomial ideal learning algorithms |
