北大经院高级计量经济学——分布理论.pdfVIP

Topic 1: Distribution Theory Advanced Econometrics (I) Dong Chen School of Economics, Peking University 1 Random Variables and Distribution Define a random variable, x, to be a variable whose value realizations are uncertain. We shall quantify this uncertainty by assigning some probability to likely values. For example, we may write Pr (x = xm ) = p, where p ∈ [0, 1] . (1.1) In (1.1), x is the name of the random variable., xm is the value that x may take, and p is the probability assigned. Random variables are either discrete (finite number of outcomes or count- able set of outcomes) or continuous (infinitely divisible set of outcomes and hence not countable). If a random variable, x, is discrete, the listing of all possible values taken by x and their associated probabilities is call the probability distribution of x, denoted by f (x). If x is continuous, we assign probabilities to intervals. In this case, define probability density function (pdf ) so that f (x) ≥ 0 and (i) Pr (a ≤ x ≤ b) = ´ b f (x) dx ≥ 0; (ii) ´ f (x) dx = 1. a x For any random variable, x, with density function f (x), the probability that x is less than or equal to some number, a, is denoted F (a), which is called the cumulative distribution function (cdf ). Formally, F (a) = ˆ a f (x) dx. (1.2) −∞ 1.1 Expectation and Variance of a Random Variable Definition 1: [Expectation] The expected value of a random variable is x x Pr (x = xm ) if x is discrete; ´ E (x) = (1.3) x xf (x) dx if x is continuous. Usually we denote