中文说明：对于一个两类分类问题，当n=100时候，用mvnrnd()函数随机产生两类样本；每一类的样本容量不小于100；2）设计最大似然估计算法对两类类条件概率密度函数进行估计；3）设计非参数估计算法对两类的类条件概率密度进行估计（任选Parzen窗法或kn-近邻法之一），并分析样本数量、窗宽、k等因素对概率密度函数估计的影响；4）分别用2）、3）中估计的类条件概率密度函数设计最小错误概率贝叶斯分类器，实现对两类样本的分类。贝叶斯判别在数据中的应用
English Description：
For a two class classification problem, when n = 100, mvnrnd() function is used to generate two classes of samples randomly; the sample size of each class is not less than 100; 2) maximum likelihood estimation algorithm is designed to estimate the conditional probability density function of two classes; 3) nonparametric estimation algorithm is designed to estimate the conditional probability density function of two classes (one of Parzen window method or kn nearest neighbor method is optional), and the results are divided into two groups This paper analyzes the influence of sample number, window width, K and other factors on the estimation of probability density function; 4) using the class conditional probability density function estimated in 2) and 3) to design the minimum error probability Bayesian classifier to realize the classification of two kinds of samples. Application of Bayesian discriminant in data processing