1
$\begingroup$

Hej, I am trying to predict the price of renting a special type of asset(time-series on a daily level). To do this I have five years' worth of data. The dataset contains of prices for different commodities which has some correlation with the renting price of the asset price I am trying to predict. I know that there is a lag. For example, oil prices 7 days ago are likely better predicting the price of my asset than oil prices of the same day.

What method should I be using to find the lag with the best correlation?

$\endgroup$

1 Answer 1

0
$\begingroup$

What method should I be using to find the lag with the best correlation?

That can be seen in the autocorrelation curve.

Or, in a broader sense, one other good heuristic would be using Persistence models, and see if which lag (or which weighted average of the last $n$ lags of the same multiple) best predicts the current day, according to your choice of error metric.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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