摘要
在运算(∧,.)和(∨,.)的基础上,提出了一类新的模糊关联存储器模型(FAMM),并证明了利用基于运算(∧,.)的FAMM处理扩张性噪声和利用基于运算(∨,.)的FAMM处理消减性噪声是非常有效的.最后,通过一种核方法对混合噪声进行处理,取得较好效果.
The author presents a novel fuzzy associative memory, modd(FAMM) based on fuzzy operation (∧,*) and ( ∨,*).FAMM with (∧,*) is robust for dilative noise and FAMM with (∨,*) is extremely robust for erosive noise. In addition, if the input patterns contain both dilative and erosive noise, indirect approach using kernel vectors is discussed.
出处
《吉首大学学报(自然科学版)》
CAS
2005年第3期111-114,126,共5页
Journal of Jishou University(Natural Sciences Edition)
关键词
关联存储器
自动关联
能力
神经网络
associative memory
autoassociative
capability
neural network