monthly data using a daily data about midasml HOT 3 CLOSED

Hi,

Yes, you can use the package with monthly/daily, weekly/daily, or even intraday data. The key is to create MIDAS structures which can be done by using mixed_freq_data & mixed_freq_data_single functions in the package. Then, you aggregate those using MIDAS polynomials (e.g. Legendre), and run sg-LASSO on stacked covariate matrix.

from midasml.

Hi Dr. Striaukas,
thank you for providing us such a neat package. I'm trying to run sg-LASSO on a set of monthly covariates to forecast a quarterly variable. I've run mixed_freq_data_single on each of the monthly covariates, but I'm still somewhat confused on how to use the lb command to Legendre-aggregate them before applying cv.sglfit. Could you kindly provide us with a simple example on how to do it?
Thank you so much,
Pedro

from midasml.

from midasml.