感知器求解逻辑分类问题方法研究  

Study on a method of using perceptron to solve logic classification problem

作　　者：刘兴华[1] 胡泽[1] 孟江[1] 

机构地区：[1]西南石油学院电信工程学院,成都610500

出　　处：《仪器仪表用户》2004年第5期4-6,共3页Instrumentation

摘　　要：本文研究在MATLAB环境下 ,利用感知器求解逻辑分类问题。首先阐述单层感知器求解逻辑“与”和逻辑“或”问题 ,其次研究了多层感知器对“异或”问题的求解。结论表明 :在MATLAB环境下 ,利用感知器对逻辑“与”、“或”、“异或”进行求解时 ,网络训练时间短 ,仿真效果好 ,而且能够实现逻辑“与”和逻辑“或”输出的 0误差。In this article we investigate logic classification problems that are solved using perceptrons on MATLAB condition.First, we illustrate using a single-layer perceptron to solve the logic AND and logic OR problems. Second, we illustrate using multilayer perceptrons to solve the logic XOR problem. The conclusion shows that, on MATLAB condition, not only time of neural networks trained is short, but also simulation results are good when using perceptrons to solve the logic AND,OR and XOR problems. At the same time, zero error output of the logic AND and the logic OR can be realized.

关 键 词：MATLAB 感知器 逻辑分类 神经网络 单层感知器 多层感知器 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]