Multiple regression time series modelling

NewUsrStat — Mon, 15 Apr 2024 13:06:39 GMT

Hi guys,

I have to run multiple regression time series modelling on seasonal infection data.

These data are: hospitalizations of patients with infections from 3 pathogens across 5 epidemic seasons. I have to model linear trend, secular trend and polynomial trend.

Specifically, what I have to model is the following (pathogens: virus1, virus2, virus3. Totally 3 pathogens):

Tier 2 variables (time trends):

Linear trend or “βs1t ”
Seasonal trend or “βs2sin(2πt/52) + βs3cos(2πt/52)”
Secular polynomial trend or “βs1t + βs2t2+βs3t3”

Tier 3 variables (time lags):

Time lag (+/- 1 weeks)
Time lag (+/- 2 weeks)
Time lag (+/- 3 weeks)

Tier 4 variables (COVID-19 indicator + interactions):

COVID-19 indicator x Weekly number of virus1 admissions
COVID-19 indicator x Weekly number of virus2 admissions
COVID-19 indicator x Weekly number of COVID-19 admissions
COVID-19 indicator x Weekly number of all other causes admissions

Is there a similar study I can inspire to and available documentation on sas coding/data preparation to deal with this predictions?

Thank you in advance

Re: Multiple regression time series modelling

ballardw — Mon, 15 Apr 2024 13:49:30 GMT

Do you have access to SAS/ETS? That module contains the main time series tools.

If you aren't sure you can run and the log will show installed modules.

proc product_status;
run;

Re: Multiple regression time series modelling

NewUsrStat — Mon, 15 Apr 2024 14:32:57 GMT

Thank you for your reply!! This is an important suggestion! I will check!

topic Multiple regression time series modelling in New SAS User

Multiple regression time series modelling

Re: Multiple regression time series modelling

Re: Multiple regression time series modelling