Course book
Machine Learning Fundamentals
Five illustrated chapters, one for each video course. Each explains the ideas in its own words and works the numbers by hand, so you can check your understanding without running any code.
- Chapter 1A First Tour of Machine LearningFeatures and labels, kinds of learning, data splits, KNN worked by hand, Naive Bayes, logistic regression and SVMs, classifier metrics, neural networks, regression, K-means and PCA.
- Chapter 2How the Classic Algorithms WorkThe normal equation and gradient descent worked by hand, regularisation, bias and variance, SVM margins and kernels, PCA steps, decision-tree impurity and ensemble methods.
- Chapter 3Building a Classifier from ScratchCollecting sketch data, feature extraction, min-max scaling worked by hand, charting the data, nearest-neighbour and KNN code, and honest evaluation.
- Chapter 4The Scikit-learn WorkflowThe estimator pattern, splitting, saving models, comparing classic algorithms, linear against logistic regression, backpropagation, overfitting, and convolutions worked by hand.
- Chapter 5Pipelines, Metrics and Meta EstimatorsPipelines and leakage, cross-validated grid search, preprocessing tools, precision and recall across thresholds worked by hand, custom scorers and meta estimators.