本实验的目的是学习和掌握两种感知器算法：批处理感知器算法和批处理裕量松弛

算法。感知器算法是通过学习两类已标记的样本，建立一个线性分类器。学习的过程就是求解感知器权系数的过程，人们通过建立一个准则函数J(a),将求解感知器权系数的问题简化为一个标量函数J(a)的极小化问题，即当a为解向量时，J(a)最小。而极小化问题常用梯度下降法来解决。本实验给出了基于梯度下降法的两种感知器算法，介绍了原理并编程实现，最后对两种算法的特点加以比较分析。-The purpose of this study is to learn and master the two kinds of perceptron algorithm: batch sensor algorithm and batch processing margin slack

Algorithm. Perceptron algorithm is by learning two kinds of labeled samples, the establishment of a linear classifier. The process of learning is the process of solving the perceptron weights, people through the establishment of a criterion function J (a), the weighting coefficients for solving perceptron problem is reduced to a minimization problem for a scalar function J (a), i.e. when a solution vector, J (a) minimum. The minimization problem commonly used gradient descent method to solve. This paper gives two kinds of perceptron algorithm based on gradient descent method, introduces the principle and program implementation, and finally compares the characteristics of the two algorithms.