B8.1 Probability, Measure and Martingales (2021-22)
Main content blocks
- Lecturer: Profile: Jan Obloj
This course develops the mathematical foundations essential for more advanced courses in probability theory. The first part of the course develops a more sophisticated understanding of measure theory and integration, first seen in Part A Integration. The second part focuses on key probabilistic concepts: independence and conditional expectation. We then introduce discrete time martingales and establish results needed to study their behaviour. This prepares the ground for continuous martingales, studied in B8.2, which are the cornerstone of stochastic calculus.
Independence of events, random variables and \(\sigma\)-algebras, relation to product measures.
The tail \(\sigma\)-algebra, Kolomogorov's 0-1 Law, lim sup and lim inf of a sequence of events, Fatou and reverse Fatou Lemma for sets, Borel-Cantelli Lemmas.
Integration and expectation, review and extension of elementary properties of the integral and convergence theorems [from Part A Integration for the Lebesgue measure on \(R\)]. Radon-Nikodym Theorem [without proof], Scheffé's Lemma. Integration on product space, Fubini/Tonelli Theorem. Different modes of convergence and their relations. Markov’s and Jensen’s inequalities. \(L^p\) spaces, Holder’s and Minkowski’s inequalities, completeness. Uniform integrability, Vitali’s convergence theorem.
Conditional expectation: definition, properties, uniquness. Conditional convergence theorems and inequalities, link with uniform integrability. Orthogonal projection in \(L^2\), existence of conditional expectation.
Filtrations and stopping times. Examples and properties. \(\sigma\)-algebra associated to a stopping time.
Martingales in discrete time: definition, examples, properties, discrete stochastic integrals. Doob’s decomposition theorem. Stopped martingales and Doob's Optional Sampling Theorem. Maximal and \(L^p\) Inequalities, Doob's Upcrossing Lemma and Martingale Convergence Theorem. Uniformly integrable martingales, convergence in \(L^1\). Backwards martingales and Kolmogorov's Strong Law of Large Numbers.
Section outline
-
Course materials include: Panopto videos, Lecture Notes, Problem Sheets, an R code, additional slides.
There are 26 videos of various length. They cover specific sections of the lecture notes and follow a uniform naming convention: video XY is the Yth video on the contents of Chapter X. All videos are accessed via Panopto and all have attached PDFs with the complete set of notes made during the video.
Lecture Notes will be updated as the term goes on - I remove typos when/if those are spotted and I try to add on more of the non-examinable material.
There are four problem sheets, see below, and each is clearly linked to the relevant chapter in the Lecture Notes and the videos.
Finally, I will post here some additional materials, e.g., R Code for SSRW examples (covered in 02 - Introduction), notes on percolation or applications further afield. Stay tuned!