regxest (MatLab R2007b)

computes the Nadaraya-Watson estimator for univariate regression. Required by XFGIBT02.m and XFGIBT03.m.

Download File

Applied Quantitative Finance

SPMtruenadloc, XFGIBT02

Click the button to demonstrate a graph view. 
Notice: This content requires Java Runtime Environment.
Java Applet and JavaScript should be allowed on your browser.

Fri, July 27 2012 by Dedy Dwi Prastyo

mh = regxest(x {,h {,K} {,v} })

mh- n x 2 or m x 2 matrix, the first column represents the sorted first column of x or the sorted v and the second column contains the regression estimate on the values of the first column.


function[mh]=regxest(x,h,K)
[n,m]=size(x);
  if m ~= 2
    disp('regxest: data matrix must contain 2 columns');
    disp('Please input the data matrix again');
    dat = input('x=');
  end
;
eh = exist('h');
  if (eh == 0)
    %h=(max(x(:,1))-min(x(:,1)))/5;
    h = 2.42*std(x(:,1))*n^(-0.2);
  end
eK = exist('K');
  if (eK == 0)
    K=1;        %quartic kernel
  else
      if K == 0 || K >4
        disp('Type of the kernel must be "1", "2", "3" or "4"');
        disp('The quartic kernel will be used');
        K = 1;
      end  
  end
;
  tmp=sortrows(x)
  x=tmp(:,1);
  y=tmp(:,2);
  %if (exist(v)==0)
  %  v=x
  % else
  %  v=sort(v)
  % endif
;
    for i = 1:n  
        s(i,1)= nw(x(i),x,y,h,K)./nw(x(i),x,ones(n,1),h,K);
    end
    %for i = 1:n  % Gaussian kernel
     %   s(i,1)= nw(x(i),x,y,5*h,K)./nw(x(i),x,ones(n,1),5*h,K)
    %end
    mh=[x s];