Announcement

Collapse
No announcement yet.
X
Collapse
  • Page of 1
  • Filter
  • Time
  • Show
Clear All
new posts
  • #1

    Degenerated Nested Logit

    Dear Statalist Community,

    I have a question regarding the estimation of a degenerated nested logit. After some research it looks this question has been previously asked, but never answered. The problem is how to estimate a degenerated nested logit model using the
    Code:
    nlogit
    command.

    I am trying to estimate the choice between studying and not studying and among those individuals that study I want to see which major they pick. The problem is that when I run the nlogit command the iteration never converges. Here I include the code, the error, and the data used:

    Code:
    nlogit major_choosen fulltime incwage || graduated: aveparinc asvab|| majornlsy:, noconstant case(CHILD_ID)
    Error:

    Code:
     nlogit major_choosen fulltime incwage || graduated: aveparinc asvab|| majornlsy:, noconstant case(CHILD_ID) 
note: 4518 cases dropped because they have only one alternative
note: 9 cases (72 obs) dropped due to no positive outcome or multiple positive outcomes per case
note: alternatives variable graduated does not vary

tree structure specified for the nested logit model

 graduated   N       majornlsy  N    k  
-----------------------------------------
 1         19984 --- 1         2498  461
                  |- 2         2498  287
                  |- 3         2498   74
                  |- 4         2498  182
                  |- 5         2498  336
                  |- 6         2498  301
                  |- 7         2498  344
                  +- 8         2498  513
-----------------------------------------
                       total  19984 2498

k = number of times alternative is chosen
N = number of observations at each level

Iteration 0:   log likelihood = -5112.1413  
Iteration 1:   log likelihood = -5112.1413  (backed up)
Iteration 2:   log likelihood = -5112.1413  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 3:   log likelihood = -5112.1413  
Iteration 4:   log likelihood = -5112.1413  (backed up)
Iteration 5:   log likelihood = -5112.1413  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 6:   log likelihood = -5112.1413  
Iteration 7:   log likelihood = -5112.1413  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 8:   log likelihood = -5112.1413  
Iteration 9:   log likelihood = -5112.1413  (backed up)
Iteration 10:  log likelihood = -5112.1413  
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 11:  log likelihood = -5112.1413  
Iteration 12:  log likelihood = -5112.1413  (backed up)
Iteration 13:  log likelihood = -5112.1413  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 14:  log likelihood = -5112.1413  (backed up)
Iteration 15:  log likelihood = -5112.1413  (backed up)
Iteration 16:  log likelihood = -5112.1413  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 17:  log likelihood = -5112.1413  
Iteration 18:  log likelihood = -5112.1413  (backed up)
Iteration 19:  log likelihood = -5112.1413  
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 20:  log likelihood = -5112.1413  
Iteration 21:  log likelihood = -5112.1413  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 22:  log likelihood = -5112.1413  
Iteration 23:  log likelihood = -5112.1413  (backed up)
Iteration 24:  log likelihood = -5112.1413  
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 25:  log likelihood = -5112.1413  (backed up)
Iteration 26:  log likelihood = -5112.1413  (backed up)
Iteration 27:  log likelihood = -5112.1413  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 28:  log likelihood = -5112.1413  (backed up)
cannot compute an improvement -- flat region encountered

    Data:


    Code:
    * Example generated by -dataex-. To install: ssc install dataex
clear
input int CHILD_ID float(aveparinc majornlsy) long asvab float(major_choosen graduated) double incwage float fulltime
6442   42468.9 0  69594 1 0 2477947.6363109956 .6811227
1363  39041.33 0      . 0 0 2475614.6363109956      .76
5337  33880.56 0  51223 0 0 2477947.6363109956 .6811227
4682   61228.8 0   4327 0 0 2477947.6363109956 .6811227
1901  63756.65 0  22738 0 0 2475614.6363109956      .76
4913 14391.074 0  84222 0 0 2477947.6363109956 .6811227
2826  28034.84 0   4301 0 0 2477947.6363109956 .6811227
3706  207782.1 0  54640 0 0 2477947.6363109956 .6811227
2014   85208.3 0      . 0 0 2477947.6363109956 .6811227
3996 19798.475 0   9246 0 0 2477947.6363109956 .6811227
5253  69565.18 0      . 0 0 2477947.6363109956 .6811227
8507  21434.74 0   2791 1 0 2477947.6363109956 .6811227
4978 19003.559 0      . 0 0 2477947.6363109956 .6811227
2260  27118.07 0      . 0 0 2477947.6363109956 .6811227
2257  51945.52 0  68920 0 0 2477947.6363109956 .6811227
8737  94262.81 0      . 1 0 2477947.6363109956 .6811227
5459 135520.44 0  20045 0 0 2477947.6363109956 .6811227
8866 23671.387 0   2469 1 0 2477947.6363109956 .6811227
7430  47726.38 0   8111 1 0 2477947.6363109956 .6811227
2475  164773.2 0  42842 0 0 2477947.6363109956 .6811227
5350  58495.14 0  43093 0 0 2477947.6363109956 .6811227
2325  65226.93 0  79063 0 0 2477947.6363109956 .6811227
 570  113373.6 0  23397 0 0 2475614.6363109956      .76
1851  56678.73 0  22172 0 0 2475614.6363109956      .76
5354  55927.88 0  83148 0 0 2477947.6363109956 .6811227
6222 111944.81 0      . 1 0 2477947.6363109956 .6811227
6106   78628.3 0  15702 1 0 2477947.6363109956 .6811227
3350  45844.54 0  31770 0 0 2477947.6363109956 .6811227
5748 35796.688 0  29464 0 0 2477947.6363109956 .6811227
5356  90411.89 0  92504 0 0 2477947.6363109956 .6811227
 830  85270.59 0      . 0 0 2475614.6363109956      .76
5860 18420.164 0  91935 0 0 2477947.6363109956 .6811227
5461  46120.61 0  16956 0 0 2477947.6363109956 .6811227
8614  38790.08 0  88622 1 0 2477947.6363109956 .6811227
8027   87894.8 0    246 1 0 2477947.6363109956 .6811227
8392 3951.8706 0  14114 1 0 2477947.6363109956 .6811227
3664 155395.27 0  27237 0 0 2477947.6363109956 .6811227
2412  84936.23 0  78371 0 0 2477947.6363109956 .6811227
7530  43385.97 0  37374 1 0 2477947.6363109956 .6811227
3209 13328.985 0      . 0 0 2477947.6363109956 .6811227
1766  49116.71 0  58786 0 0 2475614.6363109956      .76
6086  24570.23 0   3500 1 0 2477947.6363109956 .6811227
3266  81100.61 0  66661 0 0 2477947.6363109956 .6811227
3357  47456.62 0      . 0 0 2477947.6363109956 .6811227
2098 11791.878 0      . 0 0 2477947.6363109956 .6811227
 126 101610.46 0  35303 0 0 2475614.6363109956      .76
3255  74073.14 0  61634 0 0 2477947.6363109956 .6811227
5365 16137.865 0  51804 0 0 2477947.6363109956 .6811227
2809 16883.371 0  20392 0 0 2477947.6363109956 .6811227
 711 174788.55 0  70421 0 0 2475614.6363109956      .76
2122  48045.61 0  43902 0 0 2477947.6363109956 .6811227
4713  102375.9 0   1229 0 0 2477947.6363109956 .6811227
8632         . 0      . 1 0 2477947.6363109956 .6811227
1648 70232.336 0  73517 0 0 2475614.6363109956      .76
 620 11520.572 0      . 0 0 2475614.6363109956      .76
7310 11621.348 0   8004 1 0 2477947.6363109956 .6811227
3866  45132.78 0      . 0 0 2477947.6363109956 .6811227
1254  85763.32 0 100000 0 0 2475614.6363109956      .76
 160 14758.596 0   5463 0 0 2475614.6363109956      .76
7923  54932.95 0   8616 1 0 2477947.6363109956 .6811227
6777 19185.533 0  27137 1 0 2477947.6363109956 .6811227
8549  10000.21 0  12739 1 0 2477947.6363109956 .6811227
4933 22503.564 0  31111 0 0 2477947.6363109956 .6811227
1139         . 0      . 0 0 2475614.6363109956      .76
7250 12936.018 0      . 1 0 2477947.6363109956 .6811227
8563  43412.52 0  24562 1 0 2477947.6363109956 .6811227
4339   66045.8 0   9643 0 0 2477947.6363109956 .6811227
3355  48482.49 0  77827 0 0 2477947.6363109956 .6811227
9010  52539.33 0  61144 1 0 2477947.6363109956 .6811227
 866  44267.98 0      . 0 0 2475614.6363109956      .76
 460 13438.627 0   2010 0 0 2475614.6363109956      .76
6441  78058.16 0  81785 1 0 2477947.6363109956 .6811227
7580 12215.717 0   5279 1 0 2477947.6363109956 .6811227
  16 16879.293 0  44451 0 0 2475614.6363109956      .76
5414  55741.36 0  10238 0 0 2477947.6363109956 .6811227
1196  100057.4 0      . 0 0 2475614.6363109956      .76
4500 68907.555 0  23712 0 0 2477947.6363109956 .6811227
8272 14226.229 0      0 1 0 2477947.6363109956 .6811227
1833 68641.164 0  39234 0 0 2475614.6363109956      .76
2930 104045.34 0  49980 0 0 2477947.6363109956 .6811227
1347  81293.16 0  98307 0 0 2475614.6363109956      .76
5507  55387.07 0  25489 0 0 2477947.6363109956 .6811227
 582  21938.99 0   8072 0 0 2475614.6363109956      .76
1737  39426.21 0  32699 0 0 2475614.6363109956      .76
8116   81471.6 0  62873 1 0 2477947.6363109956 .6811227
  29  47374.48 0      . 0 0 2475614.6363109956      .76
3351 190708.66 0  48686 0 0 2477947.6363109956 .6811227
2502  8659.459 0   1386 0 0 2477947.6363109956 .6811227
5381  43799.02 0  69813 0 0 2477947.6363109956 .6811227
3108  21424.52 0    578 0 0 2477947.6363109956 .6811227
5382  40366.96 0  70490 0 0 2477947.6363109956 .6811227
7661  41641.71 0    135 1 0 2477947.6363109956 .6811227
4772  74203.05 0  73298 0 0 2477947.6363109956 .6811227
4773  56703.79 0  42185 0 0 2477947.6363109956 .6811227
7075  11630.74 0  14863 1 0 2477947.6363109956 .6811227
1013  90906.92 0  68139 0 0 2475614.6363109956      .76
7518 10855.753 0   9994 1 0 2477947.6363109956 .6811227
7832 11371.735 0  16329 1 0 2477947.6363109956 .6811227
5385  99861.23 0  38264 0 0 2477947.6363109956 .6811227
3942 28150.586 0      . 0 0 2477947.6363109956 .6811227
end
    Tags: convergence, degenerate, logit, nested, nlogit
Previous Next
Working...
X