[R-sig-Epi] R-sig-Epi Digest, Vol 41, Issue 1 (David Evans)
Mirian Carvalho de Souza
miriancs at inca.gov.br
Wed May 26 16:12:16 CEST 2010
Geleira
Oi David!
I'm using Poisson models.
I studied the help files of Epi package, and I didn't find a solution for
overdispersion.
I've checked and the data has overdispersion for 1-year classes and for
5-year classes L.
This is the command I'm using.
masc.fit <-
apc.fit(A=masc.df$meio,P=masc.df$periodo,D=masc.df$Y,Y=masc.df$POB/10^5,ref.c=1920,ref.p=1982,dist="poisson",model="ns",dr.extr="Holford",parm="ACP")
This is my data structure.
AGE idade meio ano PER Y POB tx_masc coorte periodo
1 1 35-39 37 1980-1984 1 400 16968583 23,573 1945 1982
2 1 35-39 37 1985-1989 2 457 20347442 22,460 1950 1987
3 1 35-39 37 1990-1994 3 453 23366339 19,387 1955 1992
4 1 35-39 37 1995-1999 4 515 26528097 19,413 1960 1997
5 1 35-39 37 2000-2004 5 423 30630477 13,810 1965 2002
6 2 40-44 42 1980-1984 1 850 15180168 55,994 1940 1982
7 2 40-44 42 1985-1989 2 940 17494477 53,731 1945 1987
8 2 40-44 42 1990-1994 3 996 19660193 50,661 1950 1992
9 2 40-44 42 1995-1999 4 1172 22532923 52,013 1955 1997
10 2 40-44 42 2000-2004 5 1221 26303422 46,420 1960 2002
11 3 45-49 47 1980-1984 1 1754 12143731 144,437 1935 1982
12 3 45-49 47 1985-1989 2 1824 13733480 132,814 1940 1987
13 3 45-49 47 1990-1994 3 1873 15293895 122,467 1945 1992
14 3 45-49 47 1995-1999 4 2261 18031010 125,395 1950 1997
15 3 45-49 47 2000-2004 5 2538 21664359 117,151 1955 2002
16 4 50-54 52 1980-1984 1 3010 10631742 283,114 1930 1982
Best regards,
Mirian Carvalho
More information about the R-sig-Epi
mailing list