mikemull / midaspy

Python version of Mixed Data Sampling (MIDAS) regression

midaspy

Python version of Mixed Data Sampling (MIDAS) regression

This is a work-in-progress. If you have cases that I can test, feel free to add an issue or a PR

midaspy's Issues

Intra-period forecasts

Hi..As in the paper,could you show how to obtain the intra-period forecasts (nowcasts) for the
gdp-payroll example.

I want to forecast a monthly series using a weekly series. Monthly series is available till July 2017 and weekly series till Aug 2017. I would like to obtain the forecasts for each week data in August 2017 for the monthly series.

Thanks

Make better results structure for estimation

Currently relying on results from scipy leastsqares optimizer. Need better output that has specific parameter names, errors, etc.

Handle nowcasting case (horizon > 0)

Test case with no high-frequency data

The case with xlag=0 should give results of a regular AR model

forecast handling multiple lags in the low-frequency variable

Hello,

Very good package. Thank you for sharing it. May I please ask if you have any plans to handle lags of level 2 or more in the adl/forecast function? When one selects ylag=2, then the result of estimate contains one more parameter estimate, which is not handled by the forecast function.

I am happy to contribute to that development.

Cyrille