list index out of range about gpfit HOT 2 CLOSED

It looks like this data set is well approximated by an affine function (straight line):

It's probably best in this case to do a simple least-squares fit, which can be accomplished by a single-term max affine fit:

In [57]: res = fit.fit(x_log, y_log, 1, "MA")
w = 0.0171 * (u_1)**3.23

The reason for the failure of the SMA fit is likely related to the fact that the data is so close to being affine. The SMA fit found above, w**424 = 0 * (u_1)**1.5e+03 + 0 * (u_1)**1.36e+03, is clearly poorly conditioned. It's fair to say gpfit should fail more gracefully in this case, but I think the simple fix is to simply use a single-term fit.

from gpfit.

Thank you!

from gpfit.