ML applications in healthcare - Saudi Arabian hospital case study
summary
Case studies use logistic regression, SVM and mention random forests applied to real hospital data and deployed in production. Logistic regression predicts one‑year mortality from ICD‑10 codes for myocardial infarction (n≈200,000), hip fracture (n≈170,000) and colorectal surgery (n≈56,000) with observed one‑year mortalities 17.2%, 27.2% and 9.3% and optimism‑adjusted C‑statistics reported as comparable. Breast cancer SVM on 10 FNA image features jumps from 34% to 98% accuracy after normalization, with one false positive in the test set, and systems are described as giving satisfactory results.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 4:15Applications of ML in healthcare
- 8:50Logistic Regression
- 10:45First case
- 13:40Breast cancer classification
- 14:50Features
- 15:40Relationship between features
- 16:30Training data
- 17:45Train model
- 18:05Data normalization
- 18:40Conclusion
related talks
33mData Science in Healthcare: Beyond the Hype
20mDiagnostic Medical application of ML
37mAI in medicine: where do we stand?
09mHow AI is changing the healthcare system
38mRealtime prediction of C-section risk for laboring mothers
23mEngineering AI-Powered Predictive Healthcare Systems
36mAI in Healthcare
54mCrossing the Last Mile with Predictable Hospital Readmissions