Machine Learning for Survival Analysis in R
Advanced Methods for Time-to-Event Modeling in Public Health & Biomedical Research
Mar 14 – Apr 26, 2026
Live via Zoom
Machine Learning for Survival Analysis in R (14-Day Weekend Program)
Duration: 14 Days (8–10 weeks)
Schedule: Saturday/Sunday, 12–2 PM ET (New York Time)
Registration Closes: March 12, 2026
Format: Live via Zoom
Payment Options: Bank Transfer, Venmo, or Zelle accepted.
Course Fee: $250 (limited scholarships and variable pricing available based on Country of participant)
About the Training
This course provides a modern, applied introduction to machine learning approaches for time-to-event (survival) data, using R and real public-health and biomedical datasets.
Participants begin with classical survival models (Kaplan–Meier, Cox proportional hazards) and progress to penalized survival models, random survival forests, gradient-boosted survival models, deep learning approaches, and explainability tools for survival ML.
The course emphasizes interpretation, diagnostics, prediction accuracy, and clear communication of findings in epidemiology, clinical research, and health policy contexts.
- Instructor-led, live Zoom sessions
- Real-world public health & biomedical datasets
- Classical → ML → Deep Learning progression
- Strong focus on interpretation & explainability
- Strong focus on interpretation & explainability
PREREQUISITES
Designed for advanced undergraduates, graduate students, faculty, and applied researchers aiming to apply ML in their work.
- Completion of Oores Analytics Machine Learning in R course
- Intermediate R skills (tidyverse, modeling basics)
- Basic probability, regression & epidemiologic concepts
- Optional: Prior exposure to ML frameworks (tidymodels, caret)
Who Should Register
Designed for advanced undergraduates, graduate students, faculty, and applied researchers aiming to apply ML in their work.
- Basic R and statistics familiarity
- No prior ML experience required
- Conceptual, not purely mathematical approach
Learning Objectives
Foundational Survival Skills
- Understand censoring, truncation, survival & hazard functions
- Fit Kaplan–Meier curves and Cox PH models
Machine Learning Skills
- Penalized Cox models (lasso, ridge, elastic net)
- Random Survival Forests & Gradient Boosting
- Deep learning survival models (DeepSurv-style)
R & Framework Skills
- Use survival, survminer, tidymodels, glmnet, ranger, gbm
- Cross-validation and tuning for censored outcomes
Interpretation & Communication
- Hazard ratios, variable importance, SHAP/LIME
- Publication-ready plots & ML explainability graphics
Weekly Breakdown
- Overview of ML concepts and workflow
- Data types, features, and labels
- Model training vs testing process
- Using real datasets for exploration (e.g., Palmer Penguins)
- Practical examples in R showing how caret and tidymodels integrate
- Handling missing values, scaling, and encoding
- Feature normalization using both caret and tidymodels
- Real dataset: Heart Disease
- Assignment: create and compare preprocessing pipelines
- Linear and multiple regression models
- Regularization (Ridge, Lasso, Elastic Net)
- Evaluation metrics (RMSE, MAE, R²)
Implementation using both caret and tidymodels
- Logistic Regression & KNN classifiers
- Model performance metrics: Accuracy, ROC, AUC, F1
- Implementation with both caret and tidymodels
Visualization with ggplot2
- Decision Trees, Random Forests, and Boosting (XGBoost)
- Bagging vs Boosting theory and implementation
- Performance evaluation and tuning
- Bonus: Feature importance visualization using vip and DALEX
- K-Means and Hierarchical Clustering
- PCA for dimensionality reduction and visualization
- Interpreting clusters and components
- Practice using cluster, factoextra, and ggplot2
- Understanding overfitting and underfitting
- K-fold and repeated cross-validation
- Grid and random search hyperparameter tuning
Building end-to-end ML pipelines with tidymodels and caret
- Feature importance and post-hoc model interpretation
- Partial Dependence Plots (PDP) and DALEX explainers
- Local explanations using LIME and SHAP
- Ethical implications and responsible ML transparency
- Deploying models using Shiny and plumber
- Model serialization with saveRDS() / readRDS()
- Building interactive prediction interfaces
- REST API creation for programmatic model access
- Best practices for versioning and reproducibility
- Full end-to-end ML workflow integration
- Real project selection & execution (classification, regression, clustering)
- Evaluation, interpretation, and optional deployment
- Deliverables: R scripts, report, and presentation
- Grading rubric & next steps for professional growth
Each week includes:
🎓 Lecture slides (PowerPoint)
💻 R lab exercises (caret + tidymodels)
📊 Real open datasets
Computing & Tools
All analyses are done in R using packages like caret, tidymodels, ggplot2, and DALEX.
All example files and syntax are downloadable from the course portal.