Skip to main content
. 2015 Nov 5;2015:953–962.

Table 3.

Notifiable cancer characteristic corpus statistics.

Category Number of Classes Majority Class Frequency Range (Mean ± Std Dev) Number of Unseen Classes in Eval.

Dev./Eval. Dev. Eval. Dev. Eval.
Basis of Diagnosis 3 08 08 19–146 (67±69) 21–175 (73±88)
Histological Type 64/94 M-81403 M-81403 1–35 (3.1±5.4) 1–21 (2.3±3.1) 65 (69%)
Histological Grade 7 9 9 3–110 (28.7±38.1) 1–129 (31.4±44.6)
Primary Site 58/66 C50.9 C42.1 1–21 (3.4±4.1) 1–39 (3.3±5.9) 30 (45%)
Laterality 4/5 8 8 20–129 (50.3±52.6) 1–134 (44.0±52.4) 1 (20%)
Metastatic Site 17/19 NA NA 1–170 (11.2±39.7) 1–192 (11.0±42.6) 10 (53%)
Metastatic Status 2 NA NA 31–170 (101±98) 28–192 (110±116)

Dev., development set (N=201); Eval., evaluation set (N=220); Std Dev, standard deviation; NA, ‘Not Applicable’ class