Hello Statalist forum users and experts,
This is my first post and I look forward to your expertise and advice.
I am trying to carry out double exponential smoothing using tssmooth on Stata 16.1 IC updated to the the August 13' 2020 update.
When allowing Stata to decide the smoothing constant alpha (tssmooth dexponential t=revtq) , I have no issues. The tsset function is used to declare panel data (tsset gvkey timer). However, when I provide a smoothing constant alpha (0.75), forecast for only the last group is displayed (tssmooth dexponential y=revtq, parms(0.75)) and the rest show up as missing data. This issue was duplicated on another machine running Stata 16.1 SE
The code used is as follows:
sort gvkey timer
tsset gvkey timer
tssmooth dexponential t=revtq
tssmooth dexponential y=revtq, parms(0.75)
As can be seen in the attached screenshot, the forecast y using smoothing coefficient alpha=0.75 is unavailable for group with gvkey=1010 and 1013, but is available for gvkey=1017.
[ATTACH=CONFIG]temp_19484_1598590965823_51[/ATTACH].
I have changed the value of alpha and the issue still remains.
Thank you in advance for taking a look and helping me out,
-Nihar
0 Response to Double exponential smoothing issues with panel data using tssmooth
Post a Comment