New answers tagged models
I don't really agree with the solution posted above. Here is why. Fitting a time series on a given stock is really trade off between statistical risk and model error. If your time serie is too short then your statistical error will be high. If your time series is too long, then the distribution of the market wil probably have changed, and the your model ...
There was a post in QF a while ago suggesting at least 5000 data points (in the time series) for a GARCH model
Top 50 recent answers are included