From what you've said, I assume you have a multi-predictor linear regression model, and that there are some indications of problematic multicollinearity.

Q1. On what basis are you concluding that you have problematic multicollinearity?

Q2. When you say Spearman method, do you mean Spearman's correlation?

If YES to Q2, given that you have a multi-predictor model, I wonder if you mean that you are considering converting all variables to their ranks. Please clarify.

EDIT: As noted in a later post, I initially misread Speakman as Spearman.

Thanks for your attention

Speakman method is one of the partial linear model estimation methods.

Dear Mr. Fras

I think that Speakman's method does not get rid of multicollinearity because it depends on partial errors in estimation

Hello Firas Monther Jassim. I am not familiar with the Spearman method for partial linear model estimation. I would be grateful if you could direct me to some introductory books, articles, or web pages that explain it. Is this article in the ballpark?

Article Covariate-adjusted Spearman's rank correlation with probabil...

Thanks.

EDIT: As noted in a later post, I initially misread Speakman as Spearman.

Apologies. I just noticed that I misread Speakman as Spearman. I'm still not familiar with it, but at least now my searches turn up some (seemingly) relevant resources. E.g.,

Article Performance of Kibria’s methods in partial linear ridge regr...

Good luck!

Some further digging makes me think you may have meant Speckman, not Speakman. ;-)

• https://stat.missouri.edu/people/speckman

Article Kernel Smoothing in Partial Linear Model

Yes Dear Bruce Weaver, Speakman's method may not be able to address the problem of multicollinearity when estimating the partial linear model, unless one of the methods of addressing the problem is used when estimating the parametric part of the said model as the principal components method

Thanks for your cooperation