Table 11 The occurrence and progression of CIN evaluated by the SCC random forest model
From: Observation of the cervical microbiome in the progression of cervical intraepithelial neoplasia
Number | Inclusion of indicators | Specific projects | Accuracy | AUC | OOB error |
|---|---|---|---|---|---|
1 | All clinical factors | age + menopause+HPV + gravidity+parity+TCT | 56.41 | 51.78 | 28.95% |
2 | Differential gene | ATM + ABCG2 + PCNA+OGG1 + TDG + LIG1 + HMGB1 | 84.61 | 84.02 | 18.42% |
3 | Differential gene + clinical factors | ATM + ABCG2 + PCNA+OGG1 + TDG + LIG1 + HMGB1 + age + menopause+HPV + TCT | 82.05 | 87.28 | 23.68% |
4 | Single-factor logistic regression P < 0.05 gene | ABCG2 + PCNA + TDG | 84.62 | 83.14 | 18.4% |
5 | The difference in single factor logistics + 0.10 gene | ABCG2 + PCNA+TDG + age + menopause+HPV + TCT | 76.92 | 83.73 | 23.68% |
6 | Multiple-factor logistic regression P < 0.05 gene | ABCG2 + TDG | 79.49 | 85.50 | 21.05% |
7 | Multiple-factor logistic regression P < 0.05 gene + different clinical factors | ABCG2 + TDG + age + menopause+HPV + TCT | 84.62 | 82.54 | 23.68% |