Convention: N = {0,1,2,...} N* = {1,2,3,...} HW problems 1.11 #1-4 (in book) The unit square, with #Omega = \infty, has probability zero for lines. Developing a single framework for discrete {heads, tails} and picking a point from [0,1]. Ex: Probability of getting a prime number choosing at random from N* Def: Let Omega be a sample space. An event space E is a non-empty collection of subsets of Omega such that: if A \in E, A^c \in E Lemma Def: iff B \in S and B \not\in A, B \in A^c if A_1,A_2,...,A_n \in E, then the union is in E. Def: An event is called *elementary* if its cardinality is 1 In general, the power set of Omega (2^Omega = \bb P(Omega)) is the collection of all subsets of omega. (sometimes called complete/total event space) Trivial event space: {empty, Omega} Take Omega = {1,..,7}. E_1 = 2^Omega. #E_1 = 2^#Omega = 2^7. E_2 = { {3}, {1,2,4,5,6,7}, empty, Omega}. This is also a valid event sp In general, regardless of cardinality of Omega, we will take E = 2^Omega because all event spaces are subsets of the power set. N, N*, Q, Z, ... are infinite countable sets. There is a bijection between these sets and N. R is an infinite uncountable set. Rigorously defining Omega = R is very difficult. But this is outside the scope of this course. Def: A probability measure is a set function, usually denoted by P, from E to [0,+\infty] = R^+ (i) P(Omega) = 1. (ii) P(Union of A_1...A_n) = \sum_i=1^n P(A_i) if A_1..A_n is pairwise disjoint. Def: Pairwise disjoint if A_i intersect A_j = empty for all i != j. Function: P: E -> [0,+\infty) For all A in E, there exists a unique P(A) Upside down A = for all Backwards E = there exists Backwards E ! = there exists a unique Review set operations, i.e. union, intersection, complementation, commutativity, distributivity, associativity For Omega finite or infinite, \sum_\Omega P({\omega_i}) = P(\Omega) = 1. P({\omega_i}) \geq 0. Uniform probabliity: P({\omega_i}) = p_i = 1/#\Omega Each element is equally likely. If A \subset \Omega, P(A) = #A/#Omega