Support Vector Machines (SVM)

Intro
1. Classification vs Regression
2. What is SVM and SVR?
3. Components: margin, support vectors, hyperplane, kernel, C, gamma.
4. Business applications

Math behind SVM and SVR - as simple as it can be!
1. The essence: support vectors, margins and hyperplanes.
2. Addressing non-linearity: kernels
3. Hyperparameters: C and Gamma
4. Regularization (memento overfitting!)
5. SVM vs SVR

Modeling steps + Python script
1. SMART business question
2. Data preparation
3. EDA and feature selection
4. Training and evaluating initial model
5. Steps if initial model doesn’t meet evaluation criteria:
- Overfitting: lower C, try simpler kernel, reduce features.
- Underfitting: increase C, try more complex kernel, add features.
- Kernel tricks for non-linear problems.
- Hyperparameter tuning: Grid Search CV.