Supervised Learning — Classification, Regression & Calibration — Hands-on

Learn how to design, train and evaluate supervised learning models that make sense for biomedical and omics workflows. Move from problem framing to algorithm choice, class imbalance handling, performance metrics and probability calibration using R and Python with reproducible pipelines.

Supervised Learning — Classification, Regression & Calibration

Help Desk · WhatsApp

Session Index

Session 1 — Supervised Learning Foundations & Problem Framing Session 2 — Classification Algorithms & Class Imbalance Session 3 — Regression Algorithms & Error Metrics Session 4 — Calibration, Thresholds & Model Comparison

Session 1

Fee: Rs 8800

Supervised Learning Foundations & Problem Framing

From clinical / omics question to ML task

classification vs regression vs ranking labels, covariates and outcomes data generating process thinking

Train/validation/test mindset

iid vs grouped data (patient, batch, site) temporal splits and leakage risks cross validation strategies

Metric families and trade offs

discrimination vs calibration ranking vs classification metrics clinical utility orientation

Session 2

Fee: Rs 11800

Classification Algorithms & Class Imbalance

Core classification algorithms

logistic regression and regularization decision trees and random forests gradient boosting (XGBoost / LightGBM)

Handling class imbalance

class weights and cost sensitive loss over/under sampling and SMOTE proper CV with resampling

Classification metrics in practice

accuracy, recall, precision, F1 ROC AUC vs PR AUC confusion matrices and imbalance

Session 3

Fee: Rs 14800

Regression Algorithms & Error Metrics

Classical and regularized regression

linear regression and assumptions Ridge / Lasso / Elastic Net non linear basis expansions

Tree based regression

regression trees random forest regressors gradient boosting regressors

Error metrics and model fit

MSE, RMSE, MAE, MAPE R² and adjusted R² residual diagnostics and heteroscedasticity

Session 4

Fee: Rs 18800

Calibration, Thresholds & Model Comparison

Probability calibration

calibration curves and reliability diagrams Platt scaling and isotonic regression Brier score

Threshold selection and decision rules

Youden index and cost curves precision/recall driven thresholds decision curve analysis concepts

Deliverables: model comparison report

side by side metric tables ROC/PR and calibration plots recommended model and threshold

PDF

Supervised Learning — Classification, Regression & Calibration Training | Biostatistics & ML for Omics

Supervised Learning — Classification, Regression & Calibration — Hands-on