Linear regression - How to do Testing and training data.

I'm trying to read from the testing and training data. But I can't make the code to do that. I'm new to linear regression. I try to implement some code from Matwork but no success yet.

clc;
 clear all;
 
 
 %x=xlsread('Data Mar-26-14 Time 1533-21.txt');
 %x = fileread('Data Mar-26-14 Time 1533-21.txt');
 myDir = uigetdir; %gets directory
 myFiles = dir(fullfile(myDir,'*.txt')); %gets all text files in struct


 %x = fileread('*.txt');    % File Name XXX.txt
 %x = textscan(data, '%f%f' , 'HeaderLines',5);    % To skip 5 Header line
 
 
 ytrain=x(:,end); % Target variable
 xtrain=zscore(x(:,1:end-1));% Normalized Predictors
 
 p=length(x);


 xtrain=[ones(length(xtrain),1) xtrain]; % one is added for calculation of biases.
 xtest=xtrain;
 ytest=ytrain;
 
 %compute cost and gradient
  iter=1000; % No. of iterations for weight updation
  
  theta=zeros(size(xtrain,2),1); % Initial weights
  
  alpha=0.1 % Learning parameter
  
  [J grad h th]=cost(theta,xtrain,ytrain,alpha,iter) % Cost funtion
 
   ypred=xtest*th; %target prediction
 
   % probability calculation
 [hp]=sigmoid(ypred); % Hypothesis Function
 ypred(hp>=0.5)=1;
 ypred(hp<0.5)=0;
 
% Decision Boundary

syms x1 x2
fnn=th(1)+th(2)*x1+th(3)*x2-0.5

 figure
 hold on
 scatter(xtest(ytest==1,2),xtest(ytest==1,3),'b+','linewidth',5.0)
 scatter(xtest(ytest==0,2),xtest(ytest==0,3),'r','linewidth',5.0)
 h1=ezplot(fnn)
 set(h1,'color','r')
 legend('Pos class','Neg. class','Decision boundary')
 xlim([-2 2])
 ylim([-2 2])
 

function fscore=cmpt_multiclass_fisher(X,y)
    uniclass = unique(y);

    
    for k = 1:2

            idxk = find( y==uniclass(k) ); nk = length(idxk);
            Xk = X(idxk,:);
            
            num(k,:) = nk.*( mean(Xk) - mean(X) ).^2;
            den(k,:) = nk.*(var(Xk) );

    end
    
    fscore = sum(num) ./ sum(den);

end