I am trying to fit a time series model on daily data for 2 years. Data is related to daily count of something. I have 731 records (as it includes leap year as well). I have been using ARIMA Model in r but the forecast are very off. Is there any specific algorithm that takes care of daily data? Also the forecast can not be in fraction as its the count data. It is always integer. Any help is appreciated!
Asked
Active
Viewed 572 times
1
-
1See https://www.statistik.tu-dortmund.de/fileadmin/user_upload/DP_0615_SFB823_Liboschik_Fokianos_Fried.pdf. But note that the tscount package is no longer on CRAN. – Rob Hyndman Jun 05 '17 at 02:08
-
@RobHyndman Thanks! How would I be able to install tscount()? – ash25 Jun 05 '17 at 03:16
-
1See https://stackoverflow.com/questions/24194409/how-do-i-install-a-package-that-has-been-archived-from-cran – Rob Hyndman Jun 05 '17 at 03:17
-
@RobHyndman Appreciate your help. Thank you very much! – ash25 Jun 05 '17 at 03:24