Linear Regression and Regularisation
Go to main | Course Page
Section 1: Logistic Regression - Level 1
- Introduction to Logistic Regression
- Download Resources
- Motivation behind Logistic Regression
- Logistic Regression Theory 1 - Sigmoid
- Logistic Regression Theory 2 - Log Odds
- Log Odds vs Sigmoid Squashing
Section 2: Logistic Regression Theory - Level 2
- Maximum Likelihood Estimation - Part 1
- Maximum Likelihood Estimation - Part 2
- Logistic Regression via Gradient Descent
- MultiClass Logistic Regression - One vs Rest
Section 3: Logistic Regression Concepts - Level 3
- Regularized Logistic Regression
- Multicollinearity - How it affectsinference
- Perturbation Technique
- Logistic Regression for Non Linearly Separable Data
- Feature Engineering Approaches
Section 4: Examining Model Fit (Project)
- Problem Statement
- EDA
3.18. Building Logistic Regression
- Perturbation Test Code Demo
- Scikit vs Statsmodels Logit API
- McFaddens Pseudo R-Squared
- Likelihood Ratio Test
- Walds Statistic
- Predicting on Test Data
- Accuracy and caveats
- Confusion Matrix, TPR, FPR, TNR, FNR
- Precision and Recall with F1 Score
- Receiver Operating Characteristic (ROC) Curve
- Concordance and Discordance
- Cross Entropy
- KS Statistic and Gains Table
- Hyper Parameter Tuning
- Evaluation Metrics Demo
- Review Questions
- Assignment