III. Enhancing Classics
III. Enhancing Classics
Let's move beyond linearity. And control the overfitting of classical models.
Let's move beyond linearity. And control the overfitting of classical models.
Nonlinearity with a Single Predictor: Spline
Generalized Additive Model: GAM
Underfitting vs. Overfitting
Ridge, LASSO and Elastic Net
Another Expression of Ridge and LASSO
Regularization
Trees are greedy; they are doing their best at every node.
Regression Tree
Classification Tree
Performance Measure of Tree Models
Pruning Trees
Bagging: Bootstrap Aggregation
Random Forest
Boosting I: AdaBoost(Adaptive Boosting)
Boosting II: Gradient Boost