Parameter optimization - lsqnonlin with normcdf

Question

berbay 2020-5-27

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/535123-parameter-optimization-lsqnonlin-with-normcdf

回答： Raunak Gupta 2020-5-30

Hi all,

I'm trying to optimize parameters with lsqnonlin to fit data. Actual data is market prices of options and I try to fit parameters to get those prices. However, since the formula is

and d1 is expressed as

and v is a function of several parameters that I'd like to optimize. Initial parameters works well but when I run lsqnonlin v goes to negative at some point and I get an error of inputs must be real for erf function in normcdf. I made that non-negative using if statement in function v but then optimization doesn't work well.

I appreciate any lead to avoid this.

Best,

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

Raunak Gupta 2020-5-30

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/535123-parameter-optimization-lsqnonlin-with-normcdf#answer_443095

Hi,

I see that the optimization variable in the question is v. Since in lsqnonlin you are mentioning v as your parameter I suggest to put a lower bound lb equal to 0 so that optimization will not return value less than zero.This can be done as mentioned in this example.

Upper bound can be mentioned as ‘inf’ because v can be any positive real number. You may remove if statement for checking the value of v because the functionality of putting bounds on optimization variable is available in lsqnonlin.