Back to Article
Exploring AI Algorithms for Cancer Classification and Prediction Using Electronic Health Records
Journal of Artificial Intelligence and Big Data
| Vol 1, Issue 1
Table 1. Comparativeresearch table for cancer classification and prediction using electronic healthrecords
| Ref | Methodology | Dataset | Result | Limitations and Future Work |
| [7] | Ensemble learning with bagging (Decision Tree, KNN) | Breast Cancer Coimbra dataset (UCI) | 100% accuracy (Decision Tree and KNN) | High accuracy may be dataset-specific; results not generalizable |
| [8] | Logistic Regression, Naïve Bayes, Decision Trees | RSI Jemursari Hospital, Surabaya (Cervical Cancer Data) | Logistic Regression achieved 95% accuracy | Limited dataset and period; focused on specific risk factors |
| [9] | Multi-instance multi-label learning | Pathologists' viewing records, slide-level annotations | Average precision of 81% (5-class), 69% (14-class) | Weakly supervised learning scenarios |
| [10] | Random Forest Classifier | Electronic Health Records (EHR) | AUC of 0.74, predictive features: medication count, age, income | Moderately accurate, limited predictive ability for follow-up |
| [11] | Adaptive ensemble voting (ANN, Logistic Algorithm) | Wisconsin Breast Cancer dataset | 98.50% accuracy (ANN with logistic algorithm) | Limited comparison with other methods; dataset reduction effects unknown |