[Quantlib-users] Least Squares Problem targetAndValue implementation

[Amine I. <ami...@gm...>]

Hi all, 

I am trying to use a classic Least Squares minimisation in QL. I came across the Least Sq cost function, and the value() method of this cost function makes a call to targetAndValue() method which is implemented in the FitAcfProblem class. I guess this has to do with the AutoCorrelationFunction implementation of Garch in the library.

I would be grateful if someone could explain to me the logic behind the design, as all I want to use is a simple Least Squares regressor to calibrate certain parameters in a model. Do I maybe have to write a customised CostFunction?

Thanks for your help,
Amine

N.B: I am using an old version of QuantLib (1.9.2)