Table 16 Logistic regression - worked example
From: Data mining of audiology patient records: factors influencing the choice of hearing aid type
Candidate variables (database record) | Actual values | Predicted log odds | Overall predicted log odds |
|---|---|---|---|
Age | 71 | Not-significant | 0 |
Gender | Male | -0.23 | -0.23 |
AC250 | 75 | 0.57 | 0.34 |
AC500 | 70 | 0.72 | 1.06 |
AC1000 | 80 | 2.08 | 3.14 |
AC2000 | 90 | 1.19 | 4.33 |
AC4000 | 100 | 0.40 | 4.73 |
AC8000 | 100 | 0.09 | 4.82 |
BC250 | 40 | -0.03 | 4.79 |
BC500 | 60 | 0.56 | 5.35 |
BC1000 | 65 | 0.56 | 5.91 |
BC2000 | 70 | 0.14 | 6.05 |
BC4000 | 70 | Not-significant | 6.05 |
Diagnosis | Tinnitus | Not-significant | 6.05 |
Hearing aid type | BTE | To be found | 6.05 |
Masker | No masker | Not-significant | 6.05 |
Mould | 2107 | 4.09 | 10.14 |
Free-text words | REV | -0.16+0.12 = -0.04 | 10.1 |