void rfit (x, y, npts, nits, sfact, nout, xout, yout) RFIT.C Level 3
float *x, *y The x and y input vectors.
int npts The number of input points.
int nits The number of iterations.
float sfact The number of points to be considered in calculating the smoothed points, represented as a fraction of npts.
int nout The number of smoothed points to be generated.
float *xout, *yout The smoothed x and y vectors.
rfit() uses weighted linear regression to calculate a smoothed set of points to scattered data. A piece-wise 
linear least square fit is made at each point. Spline smoothing is used to fill in extra points.

Generally speaking rfit() expects to return more points than were passed in. The exception is fitting a straight line to all of the input points. In this case set nout = 2 and sfact = .99.