说明：  贝叶斯分类器是一种最优分类器，它假设基于可获得的信息可以建立类别的概率密度函数，然后基于最小错误率分类准则进行分类。一般假设概率密度函数是正态分布，因为正态分布数学基础较好。问题就转化为正态分布参数的估计问题。幸运的是，参数估计问题已经被很好的解决了。 基于正态分布的协方差的不同，正态概率分布型的贝叶斯分类器可分为：欧式距离分类器、马氏距离分类器、线性判别分析、对角线性判别分析、二次判别分析和对角二次判别分析。 在具有模式的完整统计知识条件下，按照贝叶斯决策理论进行设计的一种最优分类器。分类器是对每一个输入模式赋予一个类别名称的软件或硬件装置，而贝叶斯分类器是各种分类器中分类错误概率最小或者在预先给定代价的情况下平均风险最小的分类器。它的设计方法是一种最基本的统计分类方法。 目标：编程实现正态概率分布型的贝叶斯分类器。
(Bayesian classifier is an optimal classifier. Based on the assumption that it will be the type of information can establish the probability density function, Then, based on the smallest error rate of classification criteria for classification. General assumptions probability density function is normal, because normal mathematical better foundation. The problem into normal parameters of the estimate. Fortunately, the parameter estimation problems has been very good solution. Based on the normal distribution of covariance, the normal probability distribution of the Bayesian classifier can be divided into : Continental distance classifier, Mahalanobis distance classifier, linear discriminant analysis, diagonal discriminant analysis, quadratic discriminant analysis and right angle quadratic di)