How to determine over-fitting from non linear least square optimization tool?

Question

Ajay Goyal 2018-1-8

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/376085-how-to-determine-over-fitting-from-non-linear-least-square-optimization-tool

回答： Alan Weiss 2018-1-8

I have developed a non-linear equation. It has 4 parameters to be optimized. I have trained the function with 10 experimental data using non-linear least square error optimization method. How can I determine whether my fitting is over-fitting or normal fitting?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Alan Weiss 2018-1-8

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/376085-how-to-determine-over-fitting-from-non-linear-least-square-optimization-tool#answer_299116

One typical way to do this is by cross-validation, which means fitting a subset of the data and then checking the resulting error against the remaining data for multiple subsets of the data. See, for example, Optimize a Cross-Validated SVM Classifier Using Bayesian Optimization or examples in the crossval function reference page.

Alan Weiss

MATLAB mathematical toolbox documentation