%KSMOOTHR Kernel smoother%%      W = KSMOOTHR(X,H)%% INPUT%   X    Regression dataset%   H    Width parameter (default H=1)%% OUTPUT%   W    Kernel smoother mapping%% DESCRIPTION% Train a kernel smoothing W on data X, with width parameter H.%% SEE ALSO%  KNNR, TESTR, PLOTR% Copyright: D.M.J. Tax, D.M.J.Tax@prtools.org% Faculty EWI, Delft University of Technology% P.O. Box 5031, 2600 GA Delft, The Netherlandsfunction y = ksmoothr(x,h)if nargin<2	h = 1;endif nargin<1 | isempty(x)	y = mapping(mfilename,{h});	y = setname(y,'Kernel smoother');	returnendif ~ismapping(h) %training: just store the training data	[n,d] = size(x);	W.x = +x;	W.y = gettargets(x);	W.h = h;	y = mapping(mfilename,'trained',W,1,d,1);	y = setname(y,'Kernel smoother');else	% evaluation	W = getdata(h);	[n,d] = size(x);	m = size(W.x,1);	xtst = +x;	gamma = -1/(W.h*W.h); % tiny speedup	% now go through all test data:	y = zeros(n,1);	K = exp(gamma*distm(xtst,W.x));	y = (K*W.y)./sum(K,2);	y = setdat(x,y);	end