 | My data looks like this:input outcome
ID job1-job6 time1-time6 VIRUS;
1 1001 04 72 . . . . 4 5 . . . . 1
1 1002 40 40 72 . . . 1 2 3 . . . 0
1 1003 30 40 40 . . . 4 3 4 . . . 0
1 1006 37 37 40 33 . . 4 3 3 5 . . 0
1 1007 32 . . . . . 4 . . . . . 0
0 2525 40 40 . . . . 4 5 . . . . 1
0 2526 60 40 40 34 82 35 8 5 3 5 5 2 0
0 2527 72 . . . . . 7 . . . . . 0
0 2528 40 40 . . . . 4 3 . . . . 0
0 2529 40 40 04 . . . 4 4 3 . . . 1
0 2530 35 20 40 . . . 4 5 5 . . . 1
0 2531 40 40 40 . . . 4 3 5 . . . 0
0 2744 40 . . . . . 8 . . . . . 0
Each job variable has a correspondant time variable.
I want to use do a logistic regression with ever had job and
correspondant time. Since my real data has so many job type, it's
impossible to use 'class ' command.
My problem is how can I make a new data set, which only contains
information on a certain job and its correspondant time of each obs.
|
|
