I have a set of data which need to be modelled by a function which is defined iteratively. IE $$N(t) = 2*N(t-1)+a*N(t-1)$$

Where a is the parameter to be fitted. This isn't the function, it's just as an example. How do I go about least squares when I have an iterative function? Particularly looking for an answer which I can program in MATLAB. All the toolboxes I've seen just fit a normal function like a polynomial or something.

Anyone done this before??

  • $\begingroup$ these sort of things have been done since sooo.. long time ago. $\endgroup$ – Seyhmus Güngören Oct 16 '15 at 14:35

Say you observe $N_t$ for $t=1,2,...,T$. Then you may consider the following model:$$N_t=\beta N_{t-1}+e_t$$

where $e_t$ is some error term. Under some assumptions (OLS assumptions) you may estimate $\beta$ by $$\hat{\beta}=\frac{\sum_{t=2}^TN_tN_{t-1}}{\sum_{t=2}^TN_{t-1}^2}$$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.