Looking for adaptive RLS with variable forgetting factor?

Hi Currently working on RLS Adaptive with a variable forgetting factor, wondering if someone has worked on Matlab?

lambda(.95:.98)=lambda_s; %Forgetting Factor

err(k) = mse(d,s) % Mean square error

while k=1:50

I precisely want to acheive that if my mean square error goes 20% above the certain threshold program can concider the variable Forgetting factor for better results.

Let me know if someone has worked on it.

Thanks

[Miguel Allan Garcia-Naude]

Hi there. I am trying to do the same thing in simulink by using an adaptive forgetting factor. Unfortunatel, the recursive least squares estimator does not allow you to vary the forgetting factor during the simulation:

Cannot change parameter 'Forgetting Factor: (AdaptationParameter)' of 'indirectSTR_fuzzy/Recursive Least Squares Estimator' while simulation is running

Have you managed to solve your issue?