Syllabus
Day 1: Probability spaces
- Basic probability
- Multiple events, Bayes’ rule, independence
- Random variables, expectation and variance
Day 2: Discrete and continuous distributions: a variety of commonly-used distributions.
- Coin flips and the binomial distribution
- Multinomial
- Poisson and exponential
- The standard normal: “bell curve”
- Multivariate normal. Vector-valued observations, covariances.
Day 3: Basic statistical procedures
- Dimensionality reduction: principal component analysis and SVD
- Embedding: classic multidimensional scaling
- Regression: least-squares regression and ridge regression
- Classification: Fisher linear discriminant
- Conditional probability estimation: logistic regression
Day 4: Statistical testing
- Sampling and the central limit theorem
- Hypothesis testing: null hypothesis, p-values, A/B testing
Day 5: Probabilistic models
- Naive Bayes
- Mixtures of Gaussians
- Simple Bayesian nets
- Topic models