Hello,
I’m having trouble estimating a few parameters. Basically the estimations are too close to the priors mode. I tried to decrease the number of time series and estimators, or to change the initial calibrated guess, unsuccessfully.
Thank you for looking at it!
nm04_est2.mod (12.7 KB)
calc_nm04.m (1.3 KB)
sist_nm04.m (1.7 KB)
algo042.m (1.7 KB)
Looking at the results, the moments seem to be well-matched. That seems to suggest that your prior was close to the “perfect match”.
Hey professor Pfeifer, tks for the message.
I’m not sure, but it looks for me an overfitting case. I did some adjustments,
algo042.m (1.0 KB)
nm04_est2.mod (12.6 KB)
calc_nm04.m (1.3 KB)
sist_nm04.m (1.7 KB)
runnernm.m (220 Bytes)
I took off the pandemic period sample, and I restricted the number of parameters to be estimated; the results looks more plausible now. Please, could you take a look?
What is your metric for “plausibility”?
I’m sorry I don’t have a metric, it’s just a guess.