模式识别实验：近邻法分类器设计

clear all; %导入数据 load x; load y; COL1 = size(x,2); COL2 = size(y,2); %确定k，testData，trainData，label_train的值 trainData = [x y]; a = ones(1,COL1); b = 2*ones(1,COL2); label_train = [a b]; %划分剪辑样本集和参考样本集 data_edited_x = x(:,1:(0.5*COL1)); data_edited_y = y(:,1:(0.5*COL2)); data_refer_x = x(:,(0.5*COL1+1):COL1); data_refer_y = y(:,(0.5*COL2+1):COL2); data_edited = [data_edited_x data_edited_y]; data_refer = [data_refer_x data_refer_y]; a1 = ones(1,(0.5*COL1));a2 = 2*ones(1,(0.5*COL2)); a3 = ones(1,(0.5*COL1));a4 = 2*ones(1,(0.5*COL2)); label_edited = [a1 a2]; label_refer = [a3 a4]; %对剪辑样本集进行剪辑 k1 = 5; [label_edited_test,class_edited1,class_edited2] = KNN(k1,data_refer,label_refer,data_edited); unlike = find(label_edited ~= label_edited_test); data_edited(:,unlike) = []; label_edited(:,unlike) = []; datax_edited = find(label_edited == 1); datay_edited = find(label_edited == 2); %确定k，testData，trainData，label_train的值 testData = [-0.7303 1.4445 -1.2587 1.2617 0.7302;2.1624 -0.1649 0.9187 -0.2086 1.6587]; k2_1 = 5; k2_2 = 7; k2_3 = 13; %k-近邻法判断测试点的类别 [label_test_k5,class1_k5,class2_k5] = KNN(k2_1,trainData,label_train,testData); [label_test_k7,class1_k7,class2_k7] = KNN(k2_2,trainData,label_train,testData); [label_test_k13,class1_k13,class2_k13] = KNN(k2_3,trainData,label_train,testData); %剪辑近邻法判断测试点的类别 [label_test_e5,class1_e5,class2_e5] = KNN(k2_1,data_edited,label_edited,testData); [label_test_e7,class1_e7,class2_e7] = KNN(k2_2,data_edited,label_edited,testData); [label_test_e13,class1_e13,class2_e13] = KNN(k2_3,data_edited,label_edited,testData); %—————————————————————————————————————————————————————————————— %————————————————————————图 像 绘 制 区———————————————————————— %画图初始样本集和分类结果 figure(1); subplot(2,3,1); plot(x(1,:),x(2,:),'ro'); hold on plot(y(1,:),y(2,:),'b*'); hold on plot(testData(1,class1_k5),testData(2,class1_k5),'bo'); hold on plot(testData(1,class2_k5),testData(2,class2_k5),'r*'); title('k=5时k-近邻法样本点和测试点的分布情况') subplot(2,3,2); plot(x(1,:),x(2,:),'ro'); hold on plot(y(1,:),y(2,:),'b*'); hold on plot(testData(1,class1_k7),testData(2,class1_k7),'bo'); hold on plot(testData(1,class2_k7),testData(2,class2_k7),'r*'); title('k=7时k-近邻法样本点和测试点的分布情况') subplot(2,3,3); plot(x(1,:),x(2,:),'ro'); hold on plot(y(1,:),y(2,:),'b*'); hold on plot(testData(1,class1_k13),testData(2,class1_k13),'bo'); hold on plot(testData(1,class2_k13),testData(2,class2_k13),'r*'); title('k=13时k-近邻法样本点和测试点的分布情况') %画图剪辑后样本集和分类结果 subplot(2,3,4); plot(data_edited(1,datax_edited),data_edited(2,datax_edited),'ro'); hold on plot(data_edited(1,datay_edited),data_edited(2,datay_edited),'b*'); hold on plot(testData(1,class1_e5),testData(2,class1_e5),'bo'); hold on plot(testData(1,class2_e5),testData(2,class2_e5),'r*'); title('k=5时剪辑近邻法样本点和测试点的分布情况') subplot(2,3,5); plot(data_edited(1,datax_edited),data_edited(2,datax_edited),'ro'); hold on plot(data_edited(1,datay_edited),data_edited(2,datay_edited),'b*'); hold on plot(testData(1,class1_e7),testData(2,class1_e7),'bo'); hold on plot(testData(1,class2_e7),testData(2,class2_e7),'r*'); title('k=7时剪辑近邻法样本点和测试点的分布情况') subplot(2,3,6); plot(data_edited(1,datax_edited),data_edited(2,datax_edited),'ro'); hold on plot(data_edited(1,datay_edited),data_edited(2,datay_edited),'b*'); hold on plot(testData(1,class1_e13),testData(2,class1_e13),'bo'); hold on plot(testData(1,class2_e13),testData(2,class2_e13),'r*'); legend('第一类样本点','第二类样本点','第一类测试点','第二类测试点'); title('k=13时剪辑近邻法样本点和测试点的分布情况') %—————————————————————————————————————————————————————————————— %——————————————————————————————————————————————————————————————