用matlab代码实现朴素贝叶斯分类器,要求如下:1.导入鸢尾花数据集; 2.分别取三个类中每个类的40个样本作为训练集,10个样本作为测试集; 3.计算每个类的概率密度函数,即每个类的样本均值与方差(假设样本特征数据

以下是使用 MATLAB 实现朴素贝叶斯分类器的代码，其中使用鸢尾花数据集进行训练和测试：

% 导入鸢尾花数据集
load fisheriris

% 将数据集分为三个类别
setosa = meas(1:50,:);
versicolor = meas(51:100,:);
virginica = meas(101:150,:);

% 取每个类别的前40个样本作为训练集，后10个样本作为测试集
train_setosa = setosa(1:40,:);
train_versicolor = versicolor(1:40,:);
train_virginica = virginica(1:40,:);
test_setosa = setosa(41:50,:);
test_versicolor = versicolor(41:50,:);
test_virginica = virginica(41:50,:);
train_data = [train_setosa; train_versicolor; train_virginica];
test_data = [test_setosa; test_versicolor; test_virginica];

% 计算每个类别的均值和方差
mean_setosa = mean(train_setosa);
var_setosa = var(train_setosa);
mean_versicolor = mean(train_versicolor);
var_versicolor = var(train_versicolor);
mean_virginica = mean(train_virginica);
var_virginica = var(train_virginica);

% 计算每个类别的先验概率
prior_setosa = size(train_setosa,1)/size(train_data,1);
prior_versicolor = size(train_versicolor,1)/size(train_data,1);
prior_virginica = size(train_virginica,1)/size(train_data,1);

% 对测试集进行分类
result = [];
for i = 1:size(test_data,1)
    test_sample = test_data(i,:);
    % 计算每个类别的后验概率
    posterior_setosa = prod(normpdf(test_sample, mean_setosa, sqrt(var_setosa))) * prior_setosa;
    posterior_versicolor = prod(normpdf(test_sample, mean_versicolor, sqrt(var_versicolor))) * prior_versicolor;
    posterior_virginica = prod(normpdf(test_sample, mean_virginica, sqrt(var_virginica))) * prior_virginica;
    % 选择概率最大的类别
    if (posterior_setosa > posterior_versicolor) && (posterior_setosa > posterior_virginica)
        result = [result; 'setosa'];
    elseif (posterior_versicolor > posterior_setosa) && (posterior_versicolor > posterior_virginica)
        result = [result; 'versicolor'];
    else
        result = [result; 'virginica'];
    end
end

% 计算分类准确率
correct_num = sum(strcmp(result, {'setosa';'versicolor';'virginica'}));
accuracy = correct_num / size(test_data,1);
disp(['朴素贝叶斯分类器的分类准确率为: ', num2str(accuracy)])

代码中使用了 MATLAB 内置的 `normpdf` 函数来计算高斯分布的概率密度函数。对于每个测试样本，计算它在每个类别下的后验概率，并选择后验概率最大的类别作为分类结果。最后，计算分类准确率。