Dear listers,
I just produced the weirdest ROC curve I've ever seen, and I suspect
there's a bug in roctab. The curve jumps up and down which is by
definition not possible. For your amusement and curiosity I append the
data. I am aware that as a diagnostic, x is worthless to predict y.
what I did was
insheet using roc_data.out
roctab y x, graph
Thanks,
Arnold Kester
----------- roc_data.out ------------
y x
0 7
0 81
0 7
0 36
0 7
0 207
0 7
0 6
0 2
0 19
0 94
0 213
0 66
0 60
0 130
0 131
0 30
0 111
0 18
0 9
0 210
0 16
0 87
0 30
0 154
0 18
0 8
0 212
0 106
0 108
0 176
0 87
0 207
0 23
0 50
0 20
0 22.5
0 143
0 37
0 130
0 70
0 17
0 10
0 7
0 134
0 134
0 51
0 2
0 19
1 30
1 114
1 35
1 6
1 2
1 2
1 20
1 58
1 14
1 12
1 26
1 140
1 13
1 255
1 6
1 18
1 234
1 128
1 9
1 161
1 100
1 16
1 18
1 48
1 225
1 50
1 91
1 41
1 7
1 82
1 22
1 10
1 312
1 22.5
1 6
1 145
1 122
1 228
1 9
1 6
1 34
1 85
1 24
1 11
1 93
1 57
1 178
1 134
1 7
1 128
1 88
1 95
1 119
1 2
1 14
1 171
1 216
1 79
1 96
1 17
1 6
1 8
1 6
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/