神经网络基本介绍（三）：前馈网络（上）M-P model 和单层感知机

 

 

 

上图是前馈网络基本结构，分别介绍几种前馈网络： 

M-P model，single-layer perceptron network,multi-layer perceptron network, BP(back propagation)

 

(一）M-P model

 

Proposed by McCulloch & Pitts
Consists of fixed Structure & fixed Link-Weights
Restrain Convex-Touch Weight ⇒ output is 0
Exciting Convex-Touch Weight ⇒ output is 1 when
the accumulated value ≥θ

Advantages:
Simplest Feed-forward NN
Can implement some Logical Operational Func
Disadvantages:
I/O & weight must be discrete values: 0 or 1
Link-Weight can not be adjusted dynamically

 

(二）single-layer perceptron network

supervised method可调权重

 

这里y=f(s)也可以是其他函数，根据具体情况，因为这里是用来简单分类，所以sign函数就够了。

然后计算权重，使用delta study principle。 

 

分类结果输出。

 

 

 

 

缺点：Only linear classifiable patterns can be classified by Single-Layer Perceptron

linear non-classifiable patterns cannot be classified by Single-Layer Perceptron