LDA for 2 classes only, why have 3 classes?

3 次查看（过去 30 天）

Matlaber 2019-3-5

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/448387-lda-for-2-classes-only-why-have-3-classes

Hi,

I am trying to use the original code Link here LDA link

I found strange is the dataset have 3 classes, but the plotted graph is X as input and Y as output. The graph does not do any dimensional reduction and classification.

%Load Data
data=load('iris_dataset');
X = data.irisInputs';
T = data.irisTargets';
L = vec2ind(T')';
[Y, W, lambda] = LDA2(X, L);
%Plot Results
figure;
D = size(X,2);
for d=1:D
    % Original Data
    subplot(D,2,2*d-1);
    plot(X(:,d));
    ylabel(['x_' num2str(d)]);
    if d==D
        xlabel('Sample Index');
    end
    if d==1
        title('Original Data');
    end
    grid on;
    
    % Transformed Data
    subplot(D,2,2*d);
    plot(Y(:,d));
    ylabel(['y_' num2str(d)]);
    if d==D
        xlabel('Sample Index');
    end
    if d==1
        title('LDA Output');
    end
    grid on;
    
end