Linear Regression and Regularisation

Go to main | Course Page

Section 1: Logistic Regression - Level 1

  1. Introduction to Logistic Regression
  2. Download Resources
  3. Motivation behind Logistic Regression
  4. Logistic Regression Theory 1 - Sigmoid
  5. Logistic Regression Theory 2 - Log Odds
  6. Log Odds vs Sigmoid Squashing

Section 2: Logistic Regression Theory - Level 2

  1. Maximum Likelihood Estimation - Part 1
  2. Maximum Likelihood Estimation - Part 2
  3. Logistic Regression via Gradient Descent
  4. MultiClass Logistic Regression - One vs Rest

Section 3: Logistic Regression Concepts - Level 3

  1. Regularized Logistic Regression
  2. Multicollinearity - How it affectsinference
  3. Perturbation Technique
  4. Logistic Regression for Non Linearly Separable Data
  5. Feature Engineering Approaches

Section 4: Examining Model Fit (Project)

  1. Problem Statement
  2. EDA 3.18. Building Logistic Regression
  3. Perturbation Test Code Demo
  4. Scikit vs Statsmodels Logit API
  5. McFaddens Pseudo R-Squared
  6. Likelihood Ratio Test
  7. Walds Statistic

Section 5: Evaluating Performance

  1. Predicting on Test Data
  2. Accuracy and caveats
  3. Confusion Matrix, TPR, FPR, TNR, FNR
  4. Precision and Recall with F1 Score
  5. Receiver Operating Characteristic (ROC) Curve
  6. Concordance and Discordance
  7. Cross Entropy
  8. KS Statistic and Gains Table
  9. Hyper Parameter Tuning
  10. Evaluation Metrics Demo
  11. Review Questions
  12. Assignment
Report abuse