Machine Learning I

• 1.1 Learning from data
• 1.2 Kinds of machine learning
• 1.3 The machine-learning workflow
• 1.4 Tools
• 1.5 Case studies from Nigeria

Take the Chapter 1 assessment

• 2.1 Vectors, matrices, norms and inner products
• 2.2 Matrix calculus
• 2.3 Convexity and why it matters
• 2.4 Probability, likelihood and maximum likelihood
• 2.5 Worked derivations used later in the book

Take the Chapter 2 assessment

• 3.1 Loss functions and empirical risk minimisation
• 3.2 Gradient descent
• 3.3 Batch, mini-batch and stochastic gradient descent
• 3.4 Momentum and adaptive methods
• 3.5 Implementing gradient descent in NumPy

Take the Chapter 3 assessment

• 4.1 The model and its assumptions
• 4.2 Ordinary least squares
• 4.3 The gradient-descent solution
• 4.4 Polynomial features and basis functions
• 4.5 Evaluating regression models
• 4.6 Worked example: predicting house rent in Ibadan

Take the Chapter 4 assessment

• 5.1 Overfitting and underfitting
• 5.2 The bias-variance decomposition
• 5.3 Ridge regression
• 5.4 The lasso and the elastic net
• 5.5 Using penalties correctly
• 5.6 Choosing the penalty by cross-validation

Take the Chapter 5 assessment

• 6.1 The sigmoid and the log-odds
• 6.2 Cross-entropy loss from maximum likelihood
• 6.3 Training: gradient descent and Newton's method
• 6.4 Multiclass classification
• 6.5 Decision boundaries
• 6.6 Logistic regression in practice

Take the Chapter 6 assessment

• 7.1 The confusion matrix and its metrics
• 7.2 Choosing a threshold
• 7.3 ROC and precision-recall curves
• 7.4 Class imbalance
• 7.5 Calibration
• 7.6 Case study: evaluating the loan model

Take the Chapter 7 assessment

• 8.1 Distance metrics and the curse of dimensionality
• 8.2 The k-nearest-neighbour algorithm
• 8.3 Bayes' theorem as a classifier
• 8.4 Gaussian, multinomial and Bernoulli naive Bayes
• 8.5 Worked example: spam filtering

Take the Chapter 8 assessment

• 9.1 Recursive partitioning
• 9.2 Impurity measures
• 9.3 Regression trees
• 9.4 Pruning and depth control
• 9.5 Decision trees in practice

Take the Chapter 9 assessment

• 10.1 Bagging and random forests
• 10.2 Boosting
• 10.3 XGBoost and LightGBM in practice
• 10.4 Feature importance and its pitfalls
• 10.5 Case study: ensembles for loan default

Take the Chapter 10 assessment

• 11.1 The maximum-margin classifier
• 11.2 The dual problem and support vectors
• 11.3 Soft margins and the C parameter
• 11.4 Kernels and the kernel trick
• 11.5 Support vector regression

Take the Chapter 11 assessment

• 12.1 Splitting the data
• 12.2 Transforming features
• 12.3 Feature selection
• 12.4 Avoiding data leakage

Take the Chapter 12 assessment

• 13.1 k-fold cross-validation
• 13.2 Hyperparameter search
• 13.3 Learning curves and validation curves
• 13.4 Reporting results honestly

Take the Chapter 13 assessment

• 14.1 Problem framing and baseline
• 14.2 Iteration and error analysis
• 14.3 Packaging a model for deployment
• 14.4 Model cards and documentation
• 14.5 The course project

Take the Chapter 14 assessment