Questions by topic
The same 557 questions as the question bank, grouped under the chapter each one belongs to. 510 have a worked answer.
A chapter is beginner when it uses only events, variables, samples, and tables. It is intermediate when it adds one method, such as estimation, regression, a prior, or a standard classifier. It is advanced when that method is already assumed. A question has the level of its chapter.
Probability
90 questions in this part.
Random experiments and probability
17 questions.
| Question | Level | Answer |
|---|---|---|
| Three coin tosses | Beginner | ✓ |
| Weights on the two coin spaces | Beginner | ✓ |
| Drawing from 365 balls | Beginner | ✓ |
| A lower bound on an intersection | Beginner | ✓ |
| Fifty-four distinct birthdays | Beginner | ✓ |
| Matching faces | Beginner | ✓ |
| Three cards in order | Beginner | ✓ |
| A point drawn from the unit interval | Beginner | ✓ |
| Three people, one birthday | Beginner | ✓ |
| One of each court card | Beginner | ✓ |
| Three brothers in a random split | Beginner | ✓ |
| Binary vectors with five ones | Beginner | ✓ |
| Five balls out of twenty | Beginner | ✓ |
| A one received on a noisy channel | Beginner | ✓ |
| A zero received | Beginner | ✓ |
| Information about one die | Beginner | ✓ |
| A channel with unequal inputs | Beginner | ✓ |
Random variables and distributions
27 questions.
Simulation
4 questions.
| Question | Level | Answer |
|---|---|---|
| Draws from an exponential law | Beginner | ✓ |
| Draws from a uniform interval | Beginner | ✓ |
| Rescaling a standard normal draw | Beginner | ✓ |
| A shifted exponential density | Beginner | ✓ |
Joint distributions
8 questions.
| Question | Level | Answer |
|---|---|---|
| Minimum and average of uniform draws | Intermediate | ✓ |
| A joint table | Intermediate | ✓ |
| A probability as an integral | Intermediate | ✓ |
| Distribution function of a maximum | Intermediate | ✓ |
| One weight, or six weights | Intermediate | ✓ |
| A product density | Intermediate | ✓ |
| A point in a triangle | Intermediate | ✓ |
| Signs and their sum | Intermediate | ✓ |
Functions of random variables and limit theorems
21 questions.
Markov chains
6 questions.
| Question | Level | Answer |
|---|---|---|
| Two steps of a three-state chain | Intermediate | ✓ |
| The limit of the three-state chain | Intermediate | ✓ |
| Any two-state chain | Intermediate | ✓ |
| A chain that never settles | Intermediate | ✓ |
| A walk with a floor | Intermediate | ✓ |
| Where a random walk can be | Intermediate | ✓ |
Reliability
7 questions.
| Question | Level | Answer |
|---|---|---|
| Three components, one of them essential | Intermediate | ✓ |
| Structure of a k-out-of-n system | Intermediate | ✓ |
| Pivoting the bridge | Intermediate | ✓ |
| Paths and cuts of the bridge | Intermediate | ✓ |
| A string of twenty bulbs | Intermediate | ✓ |
| Mean life as an integral of survival | Intermediate | ✓ |
| Survival from the hazard | Intermediate | ✓ |
Stochastic processes
35 questions in this part.
Stochastic processes
2 questions.
| Question | Level | Answer |
|---|---|---|
| Same marginals, different joints | Advanced | ✓ |
| Stationary increments are not a fixed level | Advanced | ✓ |
Galton–Watson branching
3 questions.
| Question | Level | Answer |
|---|---|---|
| A critical offspring law | Advanced | ✓ |
| Several ancestors | Advanced | ✓ |
| The variance recurrence | Advanced | ✓ |
Random walks and first passage
3 questions.
| Question | Level | Answer |
|---|---|---|
| A fair walk between 0 and 8 | Advanced | ✓ |
| A biased walk on four steps | Advanced | ✓ |
| Checking the duration formula | Advanced | ✓ |
Discrete-time Markov chains
3 questions.
| Question | Level | Answer |
|---|---|---|
| Two steps and the stationary law | Advanced | ✓ |
| A deterministic three-cycle | Advanced | ✓ |
| A doubly stochastic matrix | Advanced | ✓ |
Classes, recurrence, and reversibility
3 questions.
| Question | Level | Answer |
|---|---|---|
| One closed class | Advanced | ✓ |
| A walk on a path of four vertices | Advanced | ✓ |
| Birth–death balance on three states | Advanced | ✓ |
Poisson processes
3 questions.
| Question | Level | Answer |
|---|---|---|
| Counts and a first wait | Advanced | ✓ |
| Three events in two hours | Advanced | ✓ |
| Distance to the nearest point | Advanced | ✓ |
Continuous-time Markov chains
3 questions.
| Question | Level | Answer |
|---|---|---|
| A two-state chain in continuous time | Advanced | ✓ |
| Generator and jump chain | Advanced | ✓ |
| Why a generator row sums to zero | Advanced | ✓ |
Birth–death chains and queues
3 questions.
| Question | Level | Answer |
|---|---|---|
| An M/M/1 queue | Advanced | ✓ |
| An M/M/∞ system | Advanced | ✓ |
| Two machines, equal rates | Advanced | ✓ |
Probability spaces and events
3 questions.
| Question | Level | Answer |
|---|---|---|
| The algebra generated by two sets | Advanced | ✓ |
| Two series of events | Advanced | ✓ |
| Continuity from above | Advanced | ✓ |
Random variables and laws
3 questions.
| Question | Level | Answer |
|---|---|---|
| Positive and negative parts | Advanced | ✓ |
| Same margins, two joints | Advanced | ✓ |
| A function that is not a distribution function | Advanced | ✓ |
Expectation and convergence
3 questions.
| Question | Level | Answer |
|---|---|---|
| L2 convergence and probability | Advanced | ✓ |
| A spike that keeps its mean | Advanced | ✓ |
| A bound for a sample proportion | Advanced | ✓ |
A first look at extremes
3 questions.
| Question | Level | Answer |
|---|---|---|
| The expected maximum of n uniforms | Advanced | ✓ |
| Scaling an exponential maximum | Advanced | ✓ |
| A heavier sample and its maximum | Advanced | ✓ |
Statistical inference
55 questions in this part.
Statistical inference
4 questions.
| Question | Level | Answer |
|---|---|---|
| Summaries of ten cartons | Intermediate | ✓ |
| Which of these are statistics? | Intermediate | ✓ |
| The expectation of the sample variance | Intermediate | ✓ |
| A squared t variable | Intermediate | ✓ |
Estimation
6 questions.
| Question | Level | Answer |
|---|---|---|
| Normal maximum likelihood estimates | Intermediate | ✓ |
| Bias of the two variance estimators | Intermediate | ✓ |
| An exponential rate | Intermediate | ✓ |
| A Poisson mean | Intermediate | ✓ |
| A shrunken sample mean | Intermediate | ✓ |
| No uniform winner | Intermediate | ✓ |
Information and bounds
2 questions.
| Question | Level | Answer |
|---|---|---|
| Information for a Poisson mean | Advanced | ✓ |
| Where the bound does not apply | Advanced | ✓ |
Rao–Blackwell and UMVU
2 questions.
| Question | Level | Answer |
|---|---|---|
| An unbiased Poisson probability | Advanced | ✓ |
| Why the conditional expectation is better | Advanced | ✓ |
Large-sample estimation
2 questions.
| Question | Level | Answer |
|---|---|---|
| How many more observations the median needs | Advanced | ✓ |
| A probability from a Poisson mean | Advanced | ✓ |
Confidence intervals
6 questions.
| Question | Level | Answer |
|---|---|---|
| A proportion in a poll | Intermediate | ✓ |
| The interval for a normal mean | Intermediate | ✓ |
| Two standard errors | Intermediate | ✓ |
| Lost hours after a safety programme | Intermediate | ✓ |
| Checking equal variances | Intermediate | ✓ |
| Sizing a poll | Intermediate | ✓ |
Hypothesis tests
6 questions.
| Question | Level | Answer |
|---|---|---|
| A critical region for a known variance | Intermediate | ✓ |
| Power of a serve test | Intermediate | ✓ |
| How small a variance? | Intermediate | ✓ |
| Paired or two-sample? | Intermediate | ✓ |
| Tyre wear on five cars | Intermediate | ✓ |
| A one-sided sign test | Intermediate | ✓ |
Most powerful tests
2 questions.
| Question | Level | Answer |
|---|---|---|
| A critical value from the lemma | Advanced | ✓ |
| Why the two-sided test is not UMP | Advanced | ✓ |
Generalised likelihood-ratio tests
2 questions.
| Question | Level | Answer |
|---|---|---|
| The ratio depends only on the between sum of squares | Advanced | ✓ |
| One term of the expansion | Advanced | ✓ |
Chi-squared tests
4 questions.
| Question | Level | Answer |
|---|---|---|
| A 1:2:1 phenotype ratio | Intermediate | ✓ |
| Is the die fair? | Intermediate | ✓ |
| A two-by-two table is a two-proportion test | Intermediate | ✓ |
| Counting degrees of freedom | Intermediate | ✓ |
Exact tests for tables
2 questions.
| Question | Level | Answer |
|---|---|---|
| A smaller table by hand | Advanced | ✓ |
| Which hypothesis is it? | Advanced | ✓ |
Linear regression
5 questions.
| Question | Level | Answer |
|---|---|---|
| Normal equations for a straight line | Intermediate | ✓ |
| A line through the origin | Intermediate | ✓ |
| How much do the ads explain? | Intermediate | ✓ |
| Where is the prediction interval narrowest? | Intermediate | ✓ |
| The correlation test is the slope test | Intermediate | ✓ |
Analysis of variance
6 questions.
| Question | Level | Answer |
|---|---|---|
| Heart rates in four age groups | Intermediate | ✓ |
| A Tukey interval for two varieties | Intermediate | ✓ |
| Tips and coupons | Intermediate | ✓ |
| Two treatments in blocks | Intermediate | ✓ |
| A contrast among the corn varieties | Intermediate | ✓ |
| Two groups | Intermediate | ✓ |
Linear hypotheses
2 questions.
| Question | Level | Answer |
|---|---|---|
| Dropping a slope | Advanced | ✓ |
| How many restrictions? | Advanced | ✓ |
Two-factor analysis of variance
2 questions.
| Question | Level | Answer |
|---|---|---|
| Read the reference cell | Intermediate | ✓ |
| An interaction sum of squares | Intermediate | ✓ |
Predictive values of a test
2 questions.
| Question | Level | Answer |
|---|---|---|
| A rare condition | Intermediate | ✓ |
| The matching negative predictive value | Intermediate | ✓ |
Data analysis
30 questions in this part.
Studies, data, and evidence
6 questions.
| Question | Level | Answer |
|---|---|---|
| Alice’s heart-rate changes | Beginner | ✓ |
| Randomising twenty people | Beginner | ✓ |
| What a randomised trial does not need to record | Beginner | ✓ |
| Facial tumours in Tasmanian devils | Beginner | ✓ |
| A dose comparison for middle-of-the-night insomnia | Beginner | ✓ |
| Alcohol and school attendance | Beginner | ✓ |
Descriptive statistics
2 questions.
| Question | Level | Answer |
|---|---|---|
| Classifying variables in a nutrition survey | Beginner | ✓ |
| Reading a correlation | Beginner | ✓ |
Estimation examples
7 questions.
| Question | Level | Answer |
|---|---|---|
| Ten normal observations | Intermediate | ✓ |
| Household spending on fuel | Intermediate | ✓ |
| Narrowing a pilot interval | Intermediate | ✓ |
| Printer paper | Intermediate | ✓ |
| Hours of sleep | Intermediate | ✓ |
| A difference of proportions | Intermediate | ✓ |
| Variance of a fill volume | Intermediate | ✓ |
Testing examples
6 questions.
| Question | Level | Answer |
|---|---|---|
| Caffeine in a cola | Intermediate | ✓ |
| A later poll on leaving | Intermediate | ✓ |
| Steel rod diameters | Intermediate | ✓ |
| A shuffling machine | Intermediate | ✓ |
| Two proportions | Intermediate | ✓ |
| Home-ground wins | Intermediate | ✓ |
Analysis of variance examples
2 questions.
| Question | Level | Answer |
|---|---|---|
| Haemoglobin and sickle-cell type | Intermediate | ✓ |
| Five treatments, ten pairs | Intermediate | ✓ |
Regression examples
3 questions.
| Question | Level | Answer |
|---|---|---|
| Height and age of pine trees | Intermediate | ✓ |
| Birth weight and mother’s age | Intermediate | ✓ |
| Rain, food, and orange yield | Intermediate | ✓ |
Linear model examples
2 questions.
| Question | Level | Answer |
|---|---|---|
| How far sharks swim | Intermediate | ✓ |
| Wages, class, and gender | Intermediate | ✓ |
Chi-squared examples
2 questions.
| Question | Level | Answer |
|---|---|---|
| Possums in five areas | Intermediate | ✓ |
| Genotype and mating success | Intermediate | ✓ |
Using R
12 questions in this part.
Using R
1 question.
| Question | Level | Answer |
|---|---|---|
| Two fits on different rows | Beginner | ✓ |
Exploratory data analysis in R
1 question.
| Question | Level | Answer |
|---|---|---|
| What one row of the air-quality table is | Beginner | ✓ |
Probability and simulation in R
1 question.
| Question | Level | Answer |
|---|---|---|
| A point mass is not an interval probability | Beginner | ✓ |
Bootstrap and numerical methods in R
1 question.
| Question | Level | Answer |
|---|---|---|
| Which unit the bootstrap draws | Intermediate | ✓ |
Mixtures and clustering in R
1 question.
| Question | Level | Answer |
|---|---|---|
| A label that was not in the fit | Intermediate | ✓ |
Bayesian computation in R
1 question.
| Question | Level | Answer |
|---|---|---|
| The posterior mean is not the observed rate | Advanced | ✓ |
Fitted models as objects
1 question.
| Question | Level | Answer |
|---|---|---|
| An interval for the mean and an interval for a new day | Intermediate | ✓ |
Generalised linear models in R
1 question.
| Question | Level | Answer |
|---|---|---|
| What a log-link coefficient multiplies | Intermediate | ✓ |
Generalised additive models in R
1 question.
| Question | Level | Answer |
|---|---|---|
| The basis dimension is not the degrees of freedom | Intermediate | ✓ |
Repeated measures in R
1 question.
| Question | Level | Answer |
|---|---|---|
| A correlation four visits cannot have | Advanced | ✓ |
Time series in R
1 question.
| Question | Level | Answer |
|---|---|---|
| A holdout that keeps time in order | Advanced | ✓ |
Checking a calculation
1 question.
| Question | Level | Answer |
|---|---|---|
| Change one thing | Intermediate | ✓ |
Bayesian statistics
133 questions in this part.
Bayesian foundations
15 questions.
| Question | Level | Answer |
|---|---|---|
| Confidence procedures | Intermediate | – |
| What makes a hypothesis test? | Intermediate | – |
| One thousand posterior draws | Intermediate | – |
| Prediction is not estimation | Intermediate | – |
| The missing variance | Intermediate | – |
| Posterior mean under squared-error loss | Intermediate | ✓ |
| Confidence and credible intervals | Intermediate | ✓ |
| Equal-tailed and highest-density intervals | Intermediate | ✓ |
| Bayesian German tank problem | Intermediate | ✓ |
| Prior beliefs and conflicting data | Intermediate | ✓ |
| Interpreting identical interval endpoints | Intermediate | ✓ |
| Likelihood, prior and posterior | Intermediate | ✓ |
| Posterior prediction | Intermediate | ✓ |
| Two kinds of interval | Intermediate | ✓ |
| MAP and squared-error loss | Intermediate | ✓ |
The Beta–Binomial model
11 questions.
| Question | Level | Answer |
|---|---|---|
| The pseudo-count trap | Intermediate | – |
| Symmetry and context | Intermediate | – |
| A proper nonconjugate prior | Intermediate | – |
| Sequential updating | Intermediate | – |
| A future run of tosses | Intermediate | – |
| A geometric observation and a uniform prior | Intermediate | ✓ |
| A logit-normal prior for a coin | Intermediate | ✓ |
| Beta-Binomial conjugacy and prior influence | Intermediate | ✓ |
| Choosing a conjugate prior | Intermediate | ✓ |
| A nonconjugate prior and a change of variable | Intermediate | ✓ |
| Conjugacy | Intermediate | ✓ |
Conjugate analysis for normal models
19 questions.
| Question | Level | Answer |
|---|---|---|
| Four limiting checks | Intermediate | – |
| Where did the probability go? | Intermediate | – |
| Find the prior rather than recalling it | Intermediate | – |
| How many summaries? | Intermediate | – |
| Full conditionals | Intermediate | – |
| Normal–inverse-\chi^2 conjugacy | Intermediate | ✓ |
| Normal estimation and prediction | Intermediate | ✓ |
| Estimation and prediction | Intermediate | ✓ |
| Lognormal variance and conjugate inference | Intermediate | ✓ |
| Unknown Normal mean and variance: conditional versus marginal | Intermediate | ✓ |
| A Gibbs full conditional under independent priors | Intermediate | ✓ |
| Joint Normal conjugacy and exact sampling | Intermediate | ✓ |
| Normal mean, known variance | Intermediate | ✓ |
| Normal variance, known mean | Intermediate | ✓ |
| Normal mean and variance | Intermediate | ✓ |
| Jointly conjugate normal–inverse-gamma | Intermediate | ✓ |
| Lognormal variance | Intermediate | ✓ |
| Improper prior for a normal sample | Intermediate | ✓ |
| Gamma and inverse-Gamma kernels | Intermediate | ✓ |
Markov chains, Gibbs sampling, and Metropolis–Hastings
31 questions.
Categorical data and the Dirichlet distribution
7 questions.
| Question | Level | Answer |
|---|---|---|
| Beta hidden inside Dirichlet | Intermediate | – |
| Why not a multinomial prior? | Intermediate | – |
| Small concentrations | Intermediate | – |
| The simplex CLT | Intermediate | – |
| Dirichlet updating | Intermediate | – |
| Restaurant ordering methods | Intermediate | ✓ |
| Advertising responses and Dirichlet inference | Intermediate | ✓ |
Bayesian hypotheses and model comparison
15 questions.
| Question | Level | Answer |
|---|---|---|
| Who owns the boundary? | Advanced | – |
| Fruit posterior | Advanced | – |
| Probability is not action | Advanced | – |
| Make the simplex integral equal one | Advanced | – |
| Two routes to the same table | Advanced | – |
| An IQ threshold | Advanced | ✓ |
| Bayes factors and posterior odds | Advanced | ✓ |
| Evidence and posterior model probabilities | Advanced | ✓ |
| Comparing Exponential and Gamma lifetime models | Advanced | ✓ |
| Testing whether a coin is fair | Advanced | ✓ |
| Poisson flaws and posterior odds | Advanced | ✓ |
| Exponential versus gamma lifetimes | Advanced | ✓ |
| Battery lifetimes and a Bayes factor | Advanced | ✓ |
| Bayesian model selection | Advanced | ✓ |
| Evidence and model odds | Advanced | ✓ |
Exponential families and conjugacy
18 questions.
| Question | Level | Answer |
|---|---|---|
| Factorisation and data reduction | Advanced | – |
| A normal statistic | Advanced | – |
| Reading conjugacy from a kernel | Advanced | – |
| Three different “flat” Poisson priors | Advanced | – |
| Conjugacy is not guaranteed by a name | Advanced | – |
| Sufficiency and the posterior | Advanced | ✓ |
| Exponential observations parameterised by rate | Advanced | ✓ |
| Exponential observations parameterised by mean | Advanced | ✓ |
| Rayleigh conjugate prior | Advanced | ✓ |
| Gamma waiting times with an inverse-Gamma prior | Advanced | ✓ |
| Pareto likelihood and Gamma posterior | Advanced | ✓ |
| Machine downtime and a Gamma prior | Advanced | ✓ |
| Conjugate Bayesian inference for a Gamma rate | Advanced | ✓ |
| Beta downtime | Advanced | ✓ |
| Geometric success probability | Advanced | ✓ |
| Inverse-gamma waiting time | Advanced | ✓ |
| Sufficiency | Advanced | ✓ |
| Exponential family | Advanced | ✓ |
Hierarchical Bayes
9 questions.
| Question | Level | Answer |
|---|---|---|
| Pooling | Advanced | – |
| Derive the shrinkage weight | Advanced | – |
| Exchangeability is a scientific claim | Advanced | – |
| Two hyperpriors | Advanced | – |
| Sample from a grid | Advanced | – |
| Full Bayes and interval width | Advanced | – |
| Rat-tumour studies and hierarchical shrinkage | Advanced | ✓ |
| Hierarchical Poisson model | Advanced | ✓ |
| Pooling | Advanced | ✓ |
Bayesian linear regression
8 questions.
| Question | Level | Answer |
|---|---|---|
| Errors and residuals | Advanced | – |
| Normal before Student | Advanced | – |
| The inverse-Gamma shape trap | Advanced | – |
| Direct sampling or Gibbs? | Advanced | – |
| Mean response or new response? | Advanced | – |
| Independent and scaled priors | Advanced | – |
| Bayesian regression and prediction | Advanced | ✓ |
| Normal–inverse-\chi^2 regression | Advanced | ✓ |
Generalised linear models
29 questions in this part.
Introduction to generalised linear models
9 questions.
| Question | Level | Answer |
|---|---|---|
| Score, Hessian, and information identity | Intermediate | ✓ |
| Logistic regression and an adjusted effect | Intermediate | ✓ |
| Exponential regression and the score | Intermediate | ✓ |
| Negative-binomial canonical link | Intermediate | ✓ |
| The GLM components | Intermediate | ✓ |
| Links | Intermediate | ✓ |
| Score and information | Intermediate | ✓ |
| Fisher scoring and IRLS | Intermediate | ✓ |
| Logit, probit and odds | Intermediate | ✓ |
Standard errors, and models for counts
7 questions.
| Question | Level | Answer |
|---|---|---|
| Poisson regression model statement | Intermediate | ✓ |
| Interpreting Poisson regression coefficients | Intermediate | ✓ |
| Deriving the Poisson log-likelihood | Intermediate | ✓ |
| Poisson regression and overdispersion | Intermediate | ✓ |
| Poisson regression for fish species | Intermediate | ✓ |
| A log-link coefficient | Intermediate | ✓ |
| Negative-binomial variation | Intermediate | ✓ |
Underdispersion, positive responses, and quasi-likelihood
6 questions.
| Question | Level | Answer |
|---|---|---|
| Gamma regression from the exponential-family form | Intermediate | ✓ |
| Overdispersion and quasi-Poisson inference | Intermediate | ✓ |
| Gamma GLM for reaction times | Intermediate | ✓ |
| Positive continuous responses | Intermediate | ✓ |
| Quasilikelihood | Intermediate | ✓ |
| Dispersion | Intermediate | ✓ |
Categorical responses
3 questions.
| Question | Level | Answer |
|---|---|---|
| Multinomial logistic regression | Intermediate | ✓ |
| Categorical response types | Intermediate | ✓ |
| Proportional odds | Intermediate | ✓ |
Paired comparisons, and how far a fit is from the data
4 questions.
| Question | Level | Answer |
|---|---|---|
| Deviance and model adequacy | Intermediate | ✓ |
| Bradley–Terry | Intermediate | ✓ |
| Saturated model and deviance | Intermediate | ✓ |
| Significance is not adequacy | Intermediate | ✓ |
Smoothing
7 questions in this part.
Smooth means
5 questions.
| Question | Level | Answer |
|---|---|---|
| A binomial model with a smooth effect | Intermediate | ✓ |
| Changing from a logit to a probit link | Intermediate | ✓ |
| Cholesterol and heart disease | Intermediate | ✓ |
| Cubic splines and knots | Intermediate | ✓ |
| Roughness penalty | Intermediate | ✓ |
Additive smooths
2 questions.
| Question | Level | Answer |
|---|---|---|
| Testing and simplifying an additive model | Advanced | ✓ |
| GAMs and effective degrees of freedom | Advanced | ✓ |
Statistical learning
144 questions in this part.
Covariance and the multivariate normal
8 questions.
| Question | Level | Answer |
|---|---|---|
| Why a covariance is positive semi-definite | Intermediate | ✓ |
| Variances of linear combinations | Intermediate | ✓ |
| Uncorrelated but dependent | Intermediate | ✓ |
| Axes of a contour | Intermediate | ✓ |
| Normal marginals without joint normality | Intermediate | ✓ |
| The bivariate conditional | Intermediate | ✓ |
| A Schur-complement check | Intermediate | ✓ |
| Conditional normal prediction | Intermediate | ✓ |
Estimation and inference for the multivariate normal
9 questions.
| Question | Level | Answer |
|---|---|---|
| Where the n-1 comes from | Intermediate | ✓ |
| Pooling two groups | Intermediate | ✓ |
| Two factors of one covariance | Intermediate | ✓ |
| A determinant identity | Intermediate | ✓ |
| A small T^2 test | Intermediate | ✓ |
| A two-sample statistic | Intermediate | ✓ |
| Whitening and its limits | Intermediate | ✓ |
| Normal conditioning and whitening | Intermediate | ✓ |
| A transformation is not a distributional assumption | Intermediate | ✓ |
Principal component analysis
7 questions.
| Question | Level | Answer |
|---|---|---|
| A two-variable analysis | Intermediate | ✓ |
| Properties of the components | Intermediate | ✓ |
| The effect of correlation | Intermediate | ✓ |
| Principal components through the singular values | Intermediate | ✓ |
| How many components | Intermediate | ✓ |
| A low-variance predictor | Intermediate | ✓ |
| Principal components as a modelling step | Intermediate | ✓ |
K-means clustering
8 questions.
| Question | Level | Answer |
|---|---|---|
| Pairs and means | Intermediate | ✓ |
| Counting partitions | Intermediate | ✓ |
| K-means by hand | Intermediate | ✓ |
| A Voronoi boundary | Intermediate | ✓ |
| Reading the gap statistic | Intermediate | ✓ |
| The within-cluster sum cannot choose K | Intermediate | ✓ |
| What a K-means diagram must show | Intermediate | ✓ |
| Euclidean clusters and mixture components | Intermediate | ✓ |
Mixtures and the EM algorithm
7 questions.
| Question | Level | Answer |
|---|---|---|
| Mean and variance of a mixture | Intermediate | ✓ |
| An E step | Intermediate | ✓ |
| An M step | Intermediate | ✓ |
| An unbounded likelihood | Intermediate | ✓ |
| Counting parameters | Intermediate | ✓ |
| Label switching | Intermediate | ✓ |
| Mixtures versus K-means | Intermediate | ✓ |
Model selection for stochastic models
7 questions.
| Question | Level | Answer |
|---|---|---|
| Posterior odds | Intermediate | ✓ |
| BIC and AIC by hand | Intermediate | ✓ |
| The Laplace integral in one dimension | Intermediate | ✓ |
| KL divergence | Intermediate | ✓ |
| A tempting wrong turn | Intermediate | ✓ |
| Choosing model complexity | Intermediate | ✓ |
| Mixture selection and uncertainty | Intermediate | ✓ |
Hierarchical clustering and kernel density estimation
6 questions.
| Question | Level | Answer |
|---|---|---|
| Linkage by hand | Intermediate | ✓ |
| Number of merges | Intermediate | ✓ |
| Kernel density by hand | Intermediate | ✓ |
| A kernel is a density | Intermediate | ✓ |
| A tempting wrong turn | Intermediate | ✓ |
| Clustering and smoothing choices | Intermediate | ✓ |
Statistical classifiers and discriminant analysis
8 questions.
| Question | Level | Answer |
|---|---|---|
| A Bayes rule in one dimension | Intermediate | ✓ |
| Unequal costs | Intermediate | ✓ |
| LDA with unequal priors | Intermediate | ✓ |
| A quadratic boundary | Intermediate | ✓ |
| Parameter counts | Intermediate | ✓ |
| Fisher’s direction | Intermediate | ✓ |
| The fitted LDA model | Intermediate | ✓ |
| LDA and principal components | Intermediate | ✓ |
Estimating error rates
7 questions.
| Question | Level | Answer |
|---|---|---|
| A confusion matrix | Intermediate | ✓ |
| Test-set size | Intermediate | ✓ |
| The 0.632 rule | Intermediate | ✓ |
| Jackknife of the mean | Intermediate | ✓ |
| Costs of cross-validation | Intermediate | ✓ |
| A tempting wrong turn | Intermediate | ✓ |
| Separate tuning from evaluation | Intermediate | ✓ |
Logistic regression
7 questions.
| Question | Level | Answer |
|---|---|---|
| Derivative of the logistic function | Intermediate | ✓ |
| Odds ratios | Intermediate | ✓ |
| Linear discriminant analysis implies a logistic model | Intermediate | ✓ |
| Multinomial probabilities | Intermediate | ✓ |
| Complete separation | Intermediate | ✓ |
| Reading the deviance | Intermediate | ✓ |
| A logistic rule in an unfamiliar setting | Intermediate | ✓ |
Nearest neighbours
8 questions.
| Question | Level | Answer |
|---|---|---|
| Neighbours by hand | Intermediate | ✓ |
| Changing units | Intermediate | ✓ |
| Rank distance | Intermediate | ✓ |
| The factor of two | Intermediate | ✓ |
| A tempting wrong turn | Intermediate | ✓ |
| Normalisation and nearest neighbours | Intermediate | ✓ |
| Tuning a nearest-neighbour classifier | Intermediate | ✓ |
| What an accuracy table does and does not say | Intermediate | ✓ |
Classification trees
7 questions.
| Question | Level | Answer |
|---|---|---|
| Three impurities | Intermediate | ✓ |
| Which split is better? | Intermediate | ✓ |
| A tree by hand | Intermediate | ✓ |
| Affine invariance | Intermediate | ✓ |
| Penalised impurity | Intermediate | ✓ |
| A tempting wrong turn | Intermediate | ✓ |
| Decision trees and validation | Intermediate | ✓ |
Ensembles, random forests, and boosting
7 questions.
| Question | Level | Answer |
|---|---|---|
| The error decomposition | Advanced | ✓ |
| Variance of an average | Advanced | ✓ |
| Out-of-bag counts | Advanced | ✓ |
| One round of AdaBoost | Advanced | ✓ |
| The exponential-loss minimiser | Advanced | ✓ |
| A tempting wrong turn | Advanced | ✓ |
| Why an ensemble might improve a tree | Advanced | ✓ |
Neural units
7 questions.
| Question | Level | Answer |
|---|---|---|
| Derivative of the sigmoid | Advanced | ✓ |
| Logic gates | Advanced | ✓ |
| A decision boundary | Advanced | ✓ |
| Perceptron by hand | Advanced | ✓ |
| An XOR network of your own | Advanced | ✓ |
| Linear hidden units | Advanced | ✓ |
| Why XOR needs an extra layer | Advanced | ✓ |
Backpropagation
7 questions.
| Question | Level | Answer |
|---|---|---|
| Forward pass | Advanced | ✓ |
| A gradient by hand | Advanced | ✓ |
| Counting weights | Advanced | ✓ |
| Symmetric start | Advanced | ✓ |
| Delta form | Advanced | ✓ |
| Saturation | Advanced | ✓ |
| One backpropagation update | Advanced | ✓ |
Networks in practice
7 questions.
| Question | Level | Answer |
|---|---|---|
| Initial weights | Advanced | ✓ |
| Softmax gradient | Advanced | ✓ |
| Weight decay | Advanced | ✓ |
| Online versus batch | Advanced | ✓ |
| Sum of squares versus accuracy | Advanced | ✓ |
| Counting for NETtalk | Advanced | ✓ |
| Complexity and early stopping | Advanced | ✓ |
Support vector machines
10 questions.
| Question | Level | Answer |
|---|---|---|
| Distance to a hyperplane | Advanced | ✓ |
| A two-point machine | Advanced | ✓ |
| Which points matter? | Advanced | ✓ |
| Kernel as an inner product | Advanced | ✓ |
| A Gram matrix | Advanced | ✓ |
| Interpreting multipliers | Advanced | ✓ |
| A tempting wrong turn | Advanced | ✓ |
| Read the geometry | Advanced | ✓ |
| A unit-norm constraint | Advanced | ✓ |
| Nonlinear separation of XOR | Advanced | ✓ |
High-dimensional analysis
9 questions.
| Question | Level | Answer |
|---|---|---|
| Rank | Advanced | ✓ |
| Ridge makes the matrix invertible | Advanced | ✓ |
| A Bonferroni threshold | Advanced | ✓ |
| A step-up subtlety | Advanced | ✓ |
| Soft threshold | Advanced | ✓ |
| A tempting wrong turn | Advanced | ✓ |
| When ordinary linear discrimination is not defined | Advanced | ✓ |
| Principal components before linear discrimination | Advanced | ✓ |
| Selection in many dimensions | Advanced | ✓ |
Graphical models
8 questions.
| Question | Level | Answer |
|---|---|---|
| Cliques and separators | Advanced | ✓ |
| Precision to graph | Advanced | ✓ |
| A three-variable chain | Advanced | ✓ |
| Conditional independence from a partitioned covariance | Advanced | ✓ |
| How many graphs? | Advanced | ✓ |
| A tempting wrong turn | Advanced | ✓ |
| Read a Gaussian graphical model | Advanced | ✓ |
| Conditional association and sparsity | Advanced | ✓ |
Time series
9 questions in this part.
Dependence along a series
7 questions.
| Question | Level | Answer |
|---|---|---|
| Process and realisation | Advanced | ✓ |
| Two stationarity definitions | Advanced | ✓ |
| White noise is not necessarily independent | Advanced | ✓ |
| ACF and PACF | Advanced | ✓ |
| AR, MA and ARMA orders | Advanced | ✓ |
| Invertibility | Advanced | ✓ |
| ARCH and GARCH | Advanced | ✓ |
Forecasting
2 questions.
| Question | Level | Answer |
|---|---|---|
| Unit roots and ARIMA | Advanced | ✓ |
| Identification and residual checks | Advanced | ✓ |