Hi Gilles,
However, as I've raised in https://issues.apache.org/jira/browse/MATH-512 I think that "GaussianFitter" needs some refactoring. Maybe you could try to see what changes would be necessary to deal with MATH-512 and such that the upgraded "GaussianFitter" will meet with your requirements at the same time.
I went ahead and implemented a "Normal" GaussianFitter. Turns out there's not that much difference in the resulting best fit parameters, and it seems four parameters are better than three. I agree with your comments in 512. I can refactor the classes as suggested, just let me know... Also if commons-math is interested I can submit the classes for the "Normal" GaussianFitter. I thought about combining the two and I think the implementation will be much cleaner if there are two separate implementations. Maybe the one that reflects the Wikipedia definition should be called "GaussianFitter" and the current one should be Called "GaussianFitterWithHeightFactor"... or perhaps ThreeParameterGaussianFitter FourParameterGaussianFitter... Cheers, - Ole --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
