lfsmooth

 a simple interface to locfit.
 output is a vector of smoothed values, at each data point.
 all locfit options, except evaluation structures, are valid.

 Example, to smooth a time series of observations,

 t = (1:100)';
 y = 2*sin(t/10) + normrnd(0,1,100,1);
 plot(t,y,'.');
 hold on;
 plot(t,lfsmooth(t,y,'nn',0.5));
 hold off;

0001 function yhat=lfsmooth(varargin)
0002 %
0003 % a simple interface to locfit.
0004 % output is a vector of smoothed values, at each data point.
0005 % all locfit options, except evaluation structures, are valid.
0006 %
0007 % Example, to smooth a time series of observations,
0008 %
0009 % t = (1:100)';
0010 % y = 2*sin(t/10) + normrnd(0,1,100,1);
0011 % plot(t,y,'.');
0012 % hold on;
0013 % plot(t,lfsmooth(t,y,'nn',0.5));
0014 % hold off;
0015 %
0016 
0017 % Minimal input validation
0018 if nargin < 1
0019    error( 'At least one input argument required' );
0020 end
0021 
0022 fit = locfit(x,varargin{:},'module','simple');
0023 yhat = fit.fit_points.fitted_values;
0024 
0025 return;