matlab-基于LDA特征提取的Bayesian贝叶斯二分类器matlab仿真-源码

function genPlotsLDA(class1,class2,w,decisionPoint,axisOpt,Title,numSamps) %Set preliminary variables lWidth = 2.5; x_ticks = [0:2:20]; ww = [0,0;w(1),w(2)]; ww = ww*100; %off the plot %Plot the Training Data with the Fisher Vector figure() scatter(class1(:,1),class1(:,2),'b.') hold on scatter(class2(:,1),class2(:,2),'r.') hold on plot(ww(:,1),ww(:,2),'g','LineWidth',lWidth); hold off axis equal axis(axisOpt) xticks(x_ticks) box on xlabel('x_1') ylabel('x_2') legend('Class 1','Class 2','Fisher Vector','Location','southeast') title(strcat(Title,' with Fisher Vector')) %Project test data A into 1D space y1 = (w.'*class1.').'; %Eq. from Chap3Part5 Slide 23 y2 = (w.'*class2.').'; %Find angle of fisher vector angle = atan(w(2)/w(1)); %Find x_1, x_2 coordinates of projections for plottng class1Proj2D = [cos(angle).*y1,sin(angle).*y1]; class2Proj2D = [cos(angle).*y2,sin(angle).*y2]; %Plot the Training Data Projected onto Fischer Vector figure() plot(ww(:,1),ww(:,2),'g','LineWidth',lWidth); hold on scatter(class1Proj2D(1:numSamps,1),class1Proj2D(1:numSamps,2),'b*') hold on scatter(class2Proj2D(1:numSamps,1),class2Proj2D(1:numSamps,2),'r*') hold on scatter(class1(1:numSamps,1),class1(1:numSamps,2),'b.') hold on scatter(class2(1:numSamps,1),class2(1:numSamps,2),'r.') hold off axis equal axis(axisOpt) xticks(x_ticks) box on xlabel('x_1') ylabel('x_2') legend('Fisher Vector','Projected Class 1','Projected Class 2','Class 1','Class 2','Location','southeast') title(strcat(Title, ' Projected onto Fisher Vector (',num2str(2*numSamps),' Samples)')) %Find the decision line that is orthogonal to decision point decisionLine = [decisionPoint*cos(angle)+((decisionPoint*sin(angle))/(tan(pi/2-angle))),0; 0,decisionPoint/(cos(pi/2-angle))]; %Scatter plot of data with fisher vector, decision point, and line figure() scatter(class1(:,1),class1(:,2),'b.') hold on scatter(class2(:,1),class2(:,2),'r.') hold on plot(ww(:,1),ww(:,2),'g','LineWidth',lWidth); hold on plot(decisionLine(:,1),decisionLine(:,2),'m','LineWidth',lWidth) hold on scatter(decisionPoint*cos(angle),decisionPoint*sin(angle),1000,'k.') hold off axis equal axis(axisOpt) xticks(x_ticks) box on xlabel('x_1') ylabel('x_2') legend('Class 1','Class 2','Fisher Vector','Decision Line','Decision Point','Location','southeast') title(strcat(Title,' with Fisher Vector, Decision Line, and Decision Point')) %Plot the data projected onto fisher vector with decision point figure() plot(ww(:,1),ww(:,2),'g','LineWidth',lWidth); hold on scatter(class1Proj2D(:,1),class1Proj2D(:,2),'b*') hold on scatter(class2Proj2D(:,1),class2Proj2D(:,2),'r*') hold on scatter(decisionPoint*cos(angle),decisionPoint*sin(angle),1000,'k.') hold off axis equal axis(axisOpt) xticks(x_ticks) box on xlabel('x_1') ylabel('x_2') legend('Fisher Vector','Projected Class 1','Projected Class 1','Decision Point','Location','southeast') title(strcat(Title, ' Projected onto Fisher Vector with Decision Point')) %Plot numSamps of data projected on fisher vector with decision point/line figure() plot(ww(:,1),ww(:,2),'g','LineWidth',lWidth); hold on plot(decisionLine(:,1),decisionLine(:,2),'m','LineWidth',lWidth) hold on scatter(class1Proj2D(1:numSamps,1),class1Proj2D(1:numSamps,2),'b*') hold on scatter(class2Proj2D(1:numSamps,1),class2Proj2D(1:numSamps,2),'r*') hold on scatter(class1(1:numSamps,1),class1(1:numSamps,2),'b.') hold on scatter(class2(1:numSamps,1),class2(1:numSamps,2),'r.') hold on scatter(decisionPoint*cos(angle),decisionPoint*sin(angle),1000,'k.') hold off axis equal axis(axisOpt) xticks(x_ticks) xlabel('x_1') ylabel('x_2') legend('Fisher Vector','Decision Line','Projected Class 1','Projected Class 2','Class 1','Class 2','Decision Point','Location','southeast') title(strcat(Title, ' Projected onto Fisher Vector with Decision Line and Point (',num2str(2*numSamps),' Samples)')) end