CHAPTER 17 / 40

Classical ML & Evaluation

7 lessons · Explore, experiment and understand.

DOJO PETYour little coding companion

DOJO / YOUR GUIDE

Let’s figure it out.

Ask me about this course or where to start.

Chapter lessons

7 lessons

17Classical ML & Evaluation

0/77 lessons

Step 76
~5m3D · 1Code · 11
★Most cinematic
Decision Trees and Information Gain
A decision tree is a sequence of axis-aligned splits that partitions feature space into rectangles. At each leaf the model predicts the majo…
Open lesson
Step 77
~5m3D · 1Code · 11
★Most cinematic
Random Forests
A random forest (Breiman 2001) is an ensemble of decision trees that vote for the final prediction. Two sources of diversity: bagging (each…
Open lesson
Step 78
~5m3D · 1Code · 11
★Most cinematic
Gradient Boosting
Gradient boosting (Friedman 2001) builds an additive model where each is a small regression tree fit to the negative gradient of the loss ev…
Open lesson
Step 79
~5m3D · 1Code · 11
★Most cinematic
Clustering: K-Means, Hierarchical, DBSCAN
Clustering is unsupervised — we group points without labels. K-Means alternates two steps until convergence: (1) assign each point to its ne…
Open lesson
Step 80
~5m3D · 1Code · 11
★Most cinematic
t-SNE and UMAP: Nonlinear Dimensionality Reduction
Linear methods like PCA preserve global structure but flatten nonlinear manifolds. t-SNE (van der Maaten & Hinton 2008) and UMAP (McInnes 20…
Open lesson
Step 81
~5mCode · 11
🎯Quick win
Classification Metrics: ROC, AUC, F1, Calibration
Accuracy is the wrong default for imbalanced classification. A model that always predicts 'majority class' gets 99% accuracy on a 99/1 datas…
Open lesson
Step 82
~5mCode · 11
🎯Quick win
Bootstrap and Confidence Intervals
An ML metric like 'test accuracy = 0.87' is a single number drawn from a distribution — if you re-sampled the test set, you'd get something…
Open lesson

Classical ML & Evaluation

Let’s figure it out.

Decision Trees and Information Gain

Random Forests

Gradient Boosting

Clustering: K-Means, Hierarchical, DBSCAN

t-SNE and UMAP: Nonlinear Dimensionality Reduction

Classification Metrics: ROC, AUC, F1, Calibration

Bootstrap and Confidence Intervals

Classical ML & Evaluation

Let’s figure it out.

Decision Trees and Information Gain

Random Forests

Gradient Boosting

Clustering: K-Means, Hierarchical, DBSCAN

t-SNE and UMAP: Nonlinear Dimensionality Reduction

Classification Metrics: ROC, AUC, F1, Calibration

Bootstrap and Confidence Intervals