节点文献
双向互联想网络抗噪声能力提高方法
A New Method to Improve Performance of BAM in presence of Noise
【摘要】 分析了双向互联想记忆网络的算法结构。其本质是一种对样本矢量的加权求和,权系数由输入矢量和样本矢量的内积确定,网络的抗噪声能力决定于权系数的形式。采取侧向抑制的权系数方法能获得最大抗噪声能力。
【Abstract】 Algorithm of bidirectional associative memory (BAM)is depicted.BAM sum up all sample pattern with different weighting factor to recall the true pattern.The ability to recall pattern correctly is determined by how the sample pa ttern are weighted.Selecting the maximum of weighting factors by competition can get best performance of BAM in presence of noise.
- 【文献出处】 中国医疗器械杂志 ,CHINESE JOURNAL OF MEDICAL INSTRUMEN TATION , 编辑部邮箱 ,2000年01期
- 【分类号】R-39
- 【下载频次】18