I am trying to run a hurdle model where I want to look at the significance of being enrolled under a public health insurance scheme on the incidence of hospitalisation and if hospitalised, then the number of days spent in the hospital.
My sample has a total of 268773 observations (includes 34207 that witnessed hospitalisation and 234566 with no incidence of hospitalisation). I start from a very basic model and run the following command:
churdle linear durationofstayinhosp Socialgroup gender whethsuffrngfromchronicail pc_hhdexp, ll(0) select(coveredunderpfhionly Sector pc_hhdexp)
{
durationofstayinhosp is the duration of days spent in the hospital
Socialgroup is the social group that the individual belongs to
whethsuffrngfromchronicail is whether a person suffers from chronic ailment
coveredunderpfhionly is if the individual is covered under a public health insurance scheme
Sector is urban or rural
pc_hhdexp is the per capita household expenditure
}
when I run the command the number of iterations are crossing 987(the last I checked)
Iteration 0: log likelihood = -2404.1472
Iteration 1: log likelihood = -2162.3572
Iteration 2: log likelihood = -2134.5394
Iteration 3: log likelihood = -2113.3283
Iteration 4: log likelihood = -2104.3355 (not concave)
Iteration 5: log likelihood = -2103.8891
Iteration 6: log likelihood = -2091.3969 (not concave)
Iteration 7: log likelihood = -2091.1294 (not concave)
Iteration 8: log likelihood = -2091.0952 (not concave)
Iteration 9: log likelihood = -2091.0952 (not concave)
Iteration 10: log likelihood = -2091.0952 (not concave)
Iteration 11: log likelihood = -2091.0952 (not concave)
Iteration 12: log likelihood = -2091.0952 (not concave)
Iteration 13: log likelihood = -2091.0952 (not concave)
Iteration 14: log likelihood = -2091.0952 (not concave)
Iteration 15: log likelihood = -2091.0952 (not concave)
Iteration 16: log likelihood = -2091.0952 (not concave)
-
-
-
Iteration 975: log likelihood = -2091.0952 (not concave)
Iteration 976: log likelihood = -2091.0952 (not concave)
Iteration 977: log likelihood = -2091.0952 (not concave)
Iteration 978: log likelihood = -2091.0952 (not concave)
Iteration 979: log likelihood = -2091.0952 (not concave)
Iteration 980: log likelihood = -2091.0952 (not concave)
Iteration 981: log likelihood = -2091.0952 (not concave)
Iteration 982: log likelihood = -2091.0952 (not concave)
Iteration 983: log likelihood = -2091.0952 (not concave)
Iteration 984: log likelihood = -2091.0952 (not concave)
Iteration 985: log likelihood = -2091.0952 (not concave)
Iteration 986: log likelihood = -2091.0952 (not concave)
Iteration 987: log likelihood = -2091.0952 (not concave)
However, when I fit a probit and negative binomial, then I am getting results.
Can someone point me where I am going wrong? Thanks.
My sample has a total of 268773 observations (includes 34207 that witnessed hospitalisation and 234566 with no incidence of hospitalisation). I start from a very basic model and run the following command:
churdle linear durationofstayinhosp Socialgroup gender whethsuffrngfromchronicail pc_hhdexp, ll(0) select(coveredunderpfhionly Sector pc_hhdexp)
{
durationofstayinhosp is the duration of days spent in the hospital
Socialgroup is the social group that the individual belongs to
whethsuffrngfromchronicail is whether a person suffers from chronic ailment
coveredunderpfhionly is if the individual is covered under a public health insurance scheme
Sector is urban or rural
pc_hhdexp is the per capita household expenditure
}
when I run the command the number of iterations are crossing 987(the last I checked)
Iteration 0: log likelihood = -2404.1472
Iteration 1: log likelihood = -2162.3572
Iteration 2: log likelihood = -2134.5394
Iteration 3: log likelihood = -2113.3283
Iteration 4: log likelihood = -2104.3355 (not concave)
Iteration 5: log likelihood = -2103.8891
Iteration 6: log likelihood = -2091.3969 (not concave)
Iteration 7: log likelihood = -2091.1294 (not concave)
Iteration 8: log likelihood = -2091.0952 (not concave)
Iteration 9: log likelihood = -2091.0952 (not concave)
Iteration 10: log likelihood = -2091.0952 (not concave)
Iteration 11: log likelihood = -2091.0952 (not concave)
Iteration 12: log likelihood = -2091.0952 (not concave)
Iteration 13: log likelihood = -2091.0952 (not concave)
Iteration 14: log likelihood = -2091.0952 (not concave)
Iteration 15: log likelihood = -2091.0952 (not concave)
Iteration 16: log likelihood = -2091.0952 (not concave)
-
-
-
Iteration 975: log likelihood = -2091.0952 (not concave)
Iteration 976: log likelihood = -2091.0952 (not concave)
Iteration 977: log likelihood = -2091.0952 (not concave)
Iteration 978: log likelihood = -2091.0952 (not concave)
Iteration 979: log likelihood = -2091.0952 (not concave)
Iteration 980: log likelihood = -2091.0952 (not concave)
Iteration 981: log likelihood = -2091.0952 (not concave)
Iteration 982: log likelihood = -2091.0952 (not concave)
Iteration 983: log likelihood = -2091.0952 (not concave)
Iteration 984: log likelihood = -2091.0952 (not concave)
Iteration 985: log likelihood = -2091.0952 (not concave)
Iteration 986: log likelihood = -2091.0952 (not concave)
Iteration 987: log likelihood = -2091.0952 (not concave)
However, when I fit a probit and negative binomial, then I am getting results.
Can someone point me where I am going wrong? Thanks.
