CleanData

CLEANDATA
 Usage:
       cleanx = CleanData(x, rnge, choice, epsilon);

 This function will take in a data set x, a specific portion of that data,
 rnge, and look for any points below (choice = 'below') or above (
 (choice = 'above') the threshold epsilon.  It will then get rid of those
 points and create linearly extrapolated points in their place.  It
 returns the cleaned data set cleanx.

 Note: This function is to be used ONLY when the user knows for a fact that
 spurious points exist in the data set.  It is not very automatic, but
 assures that only KNOWN false points will be removed from the set. In 
 order to use this, one must look at a plot of the data to determine the 
 portion to look at and the threshold to set.

0001 function cleanx = CleanData(x, rnge, choice, epsilon);
0002 
0003 %CLEANDATA
0004 % Usage:
0005 %       cleanx = CleanData(x, rnge, choice, epsilon);
0006 %
0007 % This function will take in a data set x, a specific portion of that data,
0008 % rnge, and look for any points below (choice = 'below') or above (
0009 % (choice = 'above') the threshold epsilon.  It will then get rid of those
0010 % points and create linearly extrapolated points in their place.  It
0011 % returns the cleaned data set cleanx.
0012 %
0013 % Note: This function is to be used ONLY when the user knows for a fact that
0014 % spurious points exist in the data set.  It is not very automatic, but
0015 % assures that only KNOWN false points will be removed from the set. In
0016 % order to use this, one must look at a plot of the data to determine the
0017 % portion to look at and the threshold to set.
0018 
0019 % Written by Dan Valente
0020 % 10 November 2006
0021 
0022 cleanx = x;
0023 dirtyx = x(rnge);
0024 
0025 if (strcmp(choice, 'below'))
0026     goodpoints = find(dirtyx > epsilon);
0027 elseif (strcmp(choice, 'above'))
0028     goodpoints = find(dirtyx < epsilon);
0029 end
0030 
0031 goodpoints = goodpoints + rnge(1) - 1;
0032 for i = 1:(length(goodpoints)-1)
0033     if ((goodpoints(i+1)-goodpoints(i)) ~= 1)
0034         xf = x(goodpoints(i+1));
0035         xs = x(goodpoints(i));
0036         m = (xf-xs)/(goodpoints(i+1)-goodpoints(i));
0037         count = goodpoints(i);
0038         while count < goodpoints(i+1)
0039             cleanx(count) = m*(count-goodpoints(i))+xs; 
0040             count = count+1;
0041         end
0042     end
0043 end
0044 
0045 return;