Help with solving the system of N linear equations

[Mark Loeffen]

Hi

Following on from an earlier conversation, I am solving a set of linear equations with 9 variables (intercept + 8 coefficients). I have figured out how to do this in lmfit, although I could just use many other Python libraries for linear regression, which I ahve also used.

My challenge is that I need to perform this fit multiple times on different datasets the change over time subject to the constraint that the correlation between the initial coefficient set (using linear regression for example) and subsequent coefficients using lmfit is greater than 0.95 for example. The technqiue is called parameter-free calibration enhancement (PFCE).

At this stage I cannot see how to code this correlation between the two sets of parameters (inital and current) and set it as a constraint for each optimisation. Your thoughts would be most helpful.

Kind regards

Mark