fisherlda.m

用mushrooms数据对模式识别课程讲述的各种模式分类方法[线性分类

clc;
clear;

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%读取数据，取16个特征
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
samples = textread('data2000.txt');
samples = samples(:,[1:6,9:15,19:22]);  %17列 第1列标号，16列特征

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%对样本进行归一化处理
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
[ms ns]=size(samples);
TMax=max(samples);
TMin=min(samples);
% 第一列是样本标签，从第二列开始归一化
for i=2:ns
    samples(:,i)=(samples(:,i)-TMin(i))/(TMax(i)-TMin(i));
end

for r = 1:5  %进行五次实验
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%将样本分为测试样本，第一类训练样本，第二类训练样本
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
p = randperm(2000);%对1：2000的整数随机排序
experiment_test=samples(p(1:500),:);%测试样本
exper_test=experiment_test(:,2:end);%测试样本,不带标签
experiment_train=samples(p(501:2000),:);%训练样本
index1=find(experiment_train(:,1)==1);%找到训练样本中第一类的行号
index2=find(experiment_train(:,1)==2);%找到训练样本中第二类的行号
exper_train_class1=experiment_train(index1,2:end);%训练样本里属于第一类的样本,不带标签
exper_train_class2=experiment_train(index2,2:end);%训练样本里属于第二类的样本,不带标签

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%计算各类的均值，散布矩阵，求投影方向W
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
mean1=mean(exper_train_class1); %计算第一类训练样本的均值
mean2=mean(exper_train_class2); %计算第二类训练样本的均值

%计算各类训练样本的散布矩阵，它是协方差矩阵的（m-1）倍，m是各类的训练样本数
[m1 n1]=size(exper_train_class1);
[m2 n2]=size(exper_train_class2);
S1=(m1-1).*cov(exper_train_class1);   
S2=(m2-1).*cov(exper_train_class2);  
SW=S1+S2; 

%也可用下面的方法直接算散布矩阵
% S1=zeros(n1,n1);
% S2=zeros(n2,n2);
% for i=1:m1
%     Rtemp=exper_train_class1(i,:);
%     S1=S1+(Rtemp-mean1)'*(Rtemp-mean1);%计算一类的散布矩阵
% end
% for i=1:m2
%     Rtemp=exper_train_class2(i,:);
%     S2=S2+(Rtemp-mean2)'*(Rtemp-mean2);%计算二类的散布矩阵
% end

SW=S1+S2;%计算总类内散布矩阵
W=inv(SW)*((mean1-mean2)');%计算投影方向W

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%对训练样本与测试样本进行降维
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
traing_example_lda_1=W'*exper_train_class1';
traing_example_lda_1=traing_example_lda_1';
traing_example_lda_2=W'*exper_train_class2';
traing_example_lda_2=traing_example_lda_2';
test_example_lda=W'*exper_test';
test_example_lda=test_example_lda';
mean_1=mean(traing_example_lda_1);
mean_2=mean(traing_example_lda_2);
point=(mean_1+mean_2)/2; %找到分类的阈值
%%%%%%%%%%%%%%%%%%%%%
%分类
%%%%%%%%%%%%%%%%%%%%%%
for j=1:size(exper_test,1);
    if test_example_lda(j,1)>point
        result_class(j,1)=1;
    else
        result_class(j,1)=2;
    end
end
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%分析结果
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
[correct(r,1),error(r,1),ROC(r,:)] = analyse_result(experiment_test,result_class);
end

%将结果放到一个数组中，便于观察数据
correct = correct';
error = error';
ROC = ROC';
result = [correct;error;ROC];
result = result';