I am writing here because I didn't find something on documentation. Sorry for spam.

Is there triple exponential smoothing availiable from pandas?

Comment From: wesm

Could you post a reference?

Comment From: jseabold

Also on my radar. https://github.com/statsmodels/statsmodels/issues/512

I think there's also a ML post along with the SO question.

http://stackoverflow.com/questions/12726432/using-python-for-time-series-analysis-and-forecasting http://en.wikipedia.org/wiki/Exponential_smoothing#Triple_exponential_smoothing

Comment From: ghost

Hi Kazoline,

Below is a link to all of the forecasting methods I have wrote up in python. The triple expo may not be 100% correct so I would recommend to test the code on your end.

https://www.box.com/s/f3u6ydgo19pzm14fy5ny

Files: Single expo Double expo python forecasting - This file also attempts to use a solver to choose the best alpha,beta,gamma parameters

The main goal will be to have this functionality built in Pandas and I hope the files above will aid the developers to reach that goal.

Comment From: dfrusdn

I made some code for Statsmodel below because most of the other forecasting functions are there

https://github.com/statsmodels/statsmodels/pull/1274

But I agree that it is much shorter and nicer in pandas

Comment From: ccsv

@kazoline and @wesm I have working code that I just submitted to statsmodel for Single, Double, and Triple and their variations (different trends and seasons) that works except I don't know how to calculate the forecast values for t1 and t2 so they have Nans. It can be found here https://github.com/statsmodels/statsmodels/pull/1287 any improvements would be appreciated.

Comment From: mbkupfer

@pichonz I know it's been many years! but could you please re-post those solutions. the box.com url you provided is no longer available.