mljs / curve-fitting Goto Github PK

this project is named curve-fitting which is a pretty general name. However, it exports directly a specific algorithm and the readme is clear about the fact that it is the "Levenberg Marquardt" curve fitting.
I think one of the two following things should be done for consistency:

• rename the project to include levenberg-marquardt. Probably ml-levenberg-marquardt on npm, with curve and fitting in the keywords. On github we can be more explicit: levenberg-marquardt-curve-fitting.
• change the structure to allow to add more algorithms in the future

I personally prefer the rename approach.

If the parameters are not passed, the default values generate some errors.

For weight, p_min, p_max, etc

The main issue is with the error computation (I think), which has to be generalized to N dimensions. It is this line that makes no sense if the data has N dimensions:

Additionally, this parameter is not documented.