*
* Example 16.6.3 from page 440
*
open data olympics.dat
data(format=free,org=columns) 1 1610 country year gdp pop gold silver bronze medaltot host planned soviet
*
set logpop = log(pop)
set loggdp = log(gdp)
*
* The Poisson model is done with DDV with the option TYPE=COUNT. The data set
* includes data for more than one Olympics, so we need to restrict ourselves to
* the sample where year is 88.
*
ddv(type=count,smpl=(year==88)) medaltot
# constant logpop loggdp
*
* The PRJPoisson procedure computes the predicted values for the Poisson model at
* the input set of X's.
*
@prjpoisson(title="Prediction at Medians",xvector=||1.0,log(5921270),log(5.51e+09)||)
@prjpoisson(title="Prediction at Higher GDP",xvector=||1.0,log(5921270),log(5.18e+10)||)
@prjpoisson(title="Prediction for UK",xvector=||1.0,log(5.72e+7),log(1.01e+12)||)