1

I have collected a value of incoming 'results' every day over a period of 4 months. I want to check if the amount of 'results' coming in every day becomes less over time. This works well with a normal regression but in three of my datasets the 'amount of results' is not normally distributed. When i Perform a poisson log regression my parameter estimates table looks weird. It has a parameter value for every day. (one sample K-S test is non significant so the data is poisson distributed)

Could you guys help me find out what i do wrong? i have the feeling its due to the fact that the date is not categorical?

SPSS code: 

DATASET ACTIVATE DataSet4.
* Generalized Linear Models.
GENLIN AANTAL_RES BY DATE (ORDER=ASCENDING)
  /MODEL DATE INTERCEPT=YES
DISTRIBUTION=POISSON LINK=LOG
  /CRITERIA METHOD=FISHER(1) SCALE=1 COVB=MODEL MAXITERATIONS=100 MAXSTEPHALVING=5 
PCONVERGE=1E-006(ABSOLUTE) SINGULAR=1E-012 ANALYSISTYPE=3(WALD) CILEVEL=95 CITYPE=WALD 
LIKELIHOOD=FULL
  /MISSING CLASSMISSING=EXCLUDE
  /PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION (EXPONENTIATED).
kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
Pascalle-r
  • 11
  • 2
  • 1
    You are not entering day as continuous. – mdewey Feb 27 '17 at 14:36
  • Some similar posts: https://stats.stackexchange.com/questions/45119/poisson-regression-with-auto-correlated-time-series, https://stats.stackexchange.com/questions/70182/time-series-dynamic-poisson-regression, https://stats.stackexchange.com/questions/312412/poisson-regression-for-time-series-data – kjetil b halvorsen Dec 27 '19 at 17:52

0 Answers0