Linear classification LDA - project line boundary

5 次查看（过去 30 天）

Ahmed Mostfa 2016-3-19

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/274462-linear-classification-lda-project-line-boundary

评论： Swarnima Khadanga 2021-4-15

LDA two classes.pdf

How can I project the line obtained from linear discriminant analysis. I couldn't find how to make it despite it is easy task. I follow the materials provided online, but no one explain how to project the line and plot the data classes.

% Number of observations of each class
n1=size(c1,1)
n2=size(c2,1)
%Mean of each class
mu1=mean(c1)
 mu2=mean(c2)
   % Average of the mean of all classes
   mu=(mu1+mu2)/2
   % Center the data (data-mean)
   d1=c1-repmat(mu1,5,1)
   d2=c2-repmat(mu2,6,1)
   % Calculate the within class variance (SW)
   s1=d1'*d1
   s2=d2'*d2
   sw=s1+s2
   invsw=inv(sw)
   % in case of two classes only use v-direction of projection
    v=invsw*(mu1-mu2)'
    %   % if more than 2 classes calculate between class variance (SB)
     sb1=n1*(mu1-mu)'*(mu1-mu)
     sb2=n2*(mu2-mu)'*(mu2-mu)
     SB=sb1+sb2
     v=invsw*SB
   % 
   % % find eigne values and eigen vectors of the (v)
   [evec,eval]=eig(v)
   % Sort eigen vectors according to eigen values (descending order) and
   % neglect eigen vectors according to small eigen values
   % v=evec(greater eigen value)
   % or use all the eigen vectors
   % project the data of the first and second class respectively
  y2=c2*v
  y1=c1*v
   unique(c1)