Table 5 Accuracy of the gene-neutral two-stage and single classifiers.
From: Identifying the status of genetic lesions in cancer clinical trial documents using machine learning
Testing Gene | Training Genes | Two-stage Classifier | Single-stage Classifier |
|---|---|---|---|
ALK | BRAF, EGFR, KIT, KRAS, MET, PTEN, WT1 | 68.6% | 64.7% |
BRAF | ALK, EGFR, KIT, KRAS, MET, PTEN, WT1 | 85.4% | 84.6% |
EGFR | ALK, BRAF, KIT, KRAS, MET, PTEN, WT1 | 78.0% | 73.5% |
KIT | ALK, BRAF, EGFR, KRAS, MET, PTEN, WT1 | 74.5% | 70.3% |
KRAS | ALK, BRAF, EGFR, KIT, MET, PTEN, WT1 | 87.0% | 81.5% |
MET | ALK, BRAF, EGFR, KIT, KRAS, PTEN, WT1 | 73.5% | 41.5% |
PTEN | ALK, BRAF, EGFR, KIT, KRAS, MET, WT1 | 78.3% | 78.3% |
WT1 | ALK, BRAF, EGFR, KIT, KRAS, MET, PTEN | 65.0% | 55.7% |
Average | Â | 76.3% | 68.8% |