Rolling window using by while loop

Question

Nazif Çat?k 2015-5-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/216521-rolling-window-using-by-while-loop

编辑： Nazif Çat?k 2015-5-14

rolling_window.m

Hello, I am currently working on the comparison of the constructed portfolios using out-of-sample variance criteria. I am going to use rolling window procedure for the comparison. First, I choose a window over which to perform estimation. Length of estimation window let say t is smaller than n, where n is the total number of returns (n=168). I use estimation window of t=120 data points which correspond to 10 years for monthly data. Second, using the return data over the estimation window, t, I compute various portfolios. I repeat this rolling window procedure for the next month and dropping the data for the earliest month. I continue this until the end of the dataset is reached. At the end of the this process, I will have generated n-t portfolio weight vectors (which should be 48 vectors in total) for each strategy. I am trying to do this with while loop in Matlab. There is a bug in the loop, it always gives the dimension error but I could not solve the problem. You can also reach my codes below. I would be very glad for your any help. Thank you

   % initialization & downloading data
      A1=importdata('data.txt');
      n1=unique(date1)
      n=length(n1)
      p=length(return1)/n
      yy=reshape(return1, [n, p])    % reshape as a matrix
      [n,p]=size(yy)
      meanx=mean(yy)
      yyy=zeros(168,450);            % convert cell to a matrix
      for i=1:168
      for j=1:450
      yyy(i,j)=yy(i,j);
      end
      end
      i=0;
      while (i<=47.5)
      x=yy(i+1:i+120,:);
      % de-mean returns
      n=size(x,1);
      p=size(x,2);
      meanx=mean(x);
      x=x-meanx(ones(n,1),:);
      xmkt=mean(x')';
      % compute sample covariance matrix
      sample=cov([x xmkt])*(n-1)/n;
      covmkt=sample(1:p,p+1);          
      varmkt=sample(p+1,p+1);
      sample(:,p+1)=[];
      sample(p+1,:)=[];
      % compute prior (F)
      prior=covmkt*covmkt'./varmkt;
      prior(logical(eye(p)))=diag(sample);           
      if (nargin < 2 | shrink == -1) % compute shrinkage parameters
        c=norm(sample-prior,'fro')^2;               
        y=x.^2;
        pr=1/n*sum(sum(y'*y))-sum(sum(sample.^2));   
        rdiag=1/n*sum(sum(y.^2))-sum(diag(sample).^2);
        z=x.*xmkt(:,ones(1,p));
        v1=1/n*y'*z-covmkt(:,ones(1,p)).*sample;
        roff1=sum(sum(v1.*covmkt(:,ones(1,p))'))/varmkt...
              -sum(diag(v1).*covmkt)/varmkt;
        v3=1/n*z'*z-varmkt*sample;
        roff3=sum(sum(v3.*(covmkt*covmkt')))/varmkt^2 ...
              -sum(diag(v3).*covmkt.^2)/varmkt^2;
        roff=2*roff1-roff3;
        r=rdiag+roff;                   
        % compute shrinkage constant (k/n)
        k=(pr-r)/c;
        shrinkage=max(0,min(1,k/n));
        else % use specified number
        shrinkage = shrink;
        end
      % compute the estimator
      sigma=shrinkage*prior+(1-shrinkage)*sample; 
      % compute the inverse of the covariance matrix
      insigma=inv(sigma);
      % portfolio optimization
      meanp=(meanx)';
      Rmax=mean(meanp);
      M=48;
      u=ones(p,1);
      a=zeros(2,2);
      a(1,1)=u'*insigma*u;
      a(1,2)=meanp'*insigma*u;
      a(2,1)=a(1,2)
      a(2,2)=meanp'*insigma*meanp;
      d=a(1,1)*a(2,2)-a(1,2)*a(1,2);
      f=(insigma*(a(2,2)*u-a(1,2)*meanp))/d;
      g=(insigma*(-a(1,2)*u+a(1,1)*meanp))/d;
      r=Rmax;
      w=zeros(450,M);
      sigmap=zeros(M,1);
      for m=1:M
      w(:,m)=f+r(m)*g;
      sigmap(m)=sqrt(w(:,m)*sigma*w(:,m));
      end 
      s= sqrt(a(1,1)*((r - a(1,2)/a(1,1))^2)/d + 1/a(1,1));
      minrisk=sqrt(1/a(1,1));
      minreturn=a(1,2)/a(1,1);
      wminp=f +(a(1,2)/a(1,1))*g;
      sharpe= Rmax/s;
      tanrisk=sqrt(a(2,2))/a(1,2);
      tanreturn=a(2,2)/a(1,2);
      wtanp=f+(a(2,2)/a(1,2))*g; 
      i=i+1
      end