STATISTICAL INFERENCE May be divided into two major areas PARAMETER ESTIMATION HYPOTHESIS TESTING POINT ESTIMATION INTERVAL ESTIMATION CONFIDENCE MEANS IE - 2333 INTERVALS The decision making procedure about the hypothesis VARIANCES PROPORTIONS SWN 1 POINT ESTIMATION A statistic used to estimate a population parameter is called a point estimator for and is denoted by ˆ . The numerical value assumed by this statistic when evaluated for a given sample is called a point estimate for . There is a difference in the terms : ESTIMATOR and ESTIMATE is the statistic used to generate the estimate ; it is a random variable IE - 2333 is a number SWN 2 We want, the estimator to generate estimates that can be expected to be close in value to . We would like : 1. ˆ to be UNBIASED for 2. ˆ to have a small variance for large sample sizes In general, If X is a random variable with probability distribution f X ( x) or pX ( x) , characterized by the unknown parameter , and if X1, X2, . . . . Xn is a random sample of size n from X, then the statistic h X 1 , X 2 , X n is called a point estimator of . note that is a random variable, because it is a function of random variable IE - 2333 SWN 3 Definition : A point estimate of some population parameter , is a single numerical value of a statistic Definition : The point estimator is an unbiased estimator for the parameter if E() . If the estimator is not unbiased, then the difference E() is called the biased of the estimator IE - 2333 SWN 4 VARIANCE AND MEAN SQUARE ERROR OF A POINT ESTIMATOR A logical principle of estimation, when selecting among several estimator, is to chose the estimator that has minimum variance. Definition : If we consider all unbiased estimator of , the one with the smallest variance is called the minimum variance unbiased estimator (MVUE). Some times the MVUE is called the UMVUE, where the first U represents “Uniformly”, meaning “for all ” IE - 2333 SWN 5 MEAN SQUARE ERROR Definition : the mean square error of an- estimator of the parameter is defined as : MSE E 2 The mean square error can be rewritten as follows : 2 MSE E E E MSE Var bias 2 2 The mean square error is an important criterion for comparing two estimators. IE - 2333 SWN 6 Let 1 and 2 be two estimators of the parameter , and let MSE 1 and MSE 2 be the mean square error of 1 and 2 . Then the relative efficiency of 1 to 2 is defined as : MSE MSE 1 2 If this relative efficiency is less than one, we would conclude that 1 is more efficient estimator of than 2 Example : Suppose we wish to estimate the mean of a population. We have a random sample of n observations X1, X2, …..Xn and we wish to compare two possible estimator for : the sample mean X and a single observation from the sample, say, Xi, IE - 2333 SWN 7 Note, both X and Xi are unbiased estimators of ; consequently, the MSE of both estimators is simply the variance. 2 We have MSE X Var X n 2 MSE 1 1 n 2 n MSE 2 Since 1n 1 for sample size n ≥ 2, we would conclude that the sample mean is a better estimator of than a single observation Xi. IE - 2333 SWN 8 EXERCISES 1. Suppose we have a random sample of size 2n from a population denoted by X, and E(X) = and Var X = 2. 2n n Let X1 1 2n X i and X 2 i 1 1 n X i 1 i be two estimator of . Which is the better estimator of ? Explain your choice. 2. Let X1, X2, . . . , X7 denote a random sample from a population having mean and variance 2. Consider the following estimator of : X 1 X 2 ..... X 7 1 ; 7 IE - 2333 2 X1 X 6 X 4 2 2 SWN 9 (a) Is either estimator unbiased? (b) Which estimator is “best” ? 3. Suppose that 1, 2 and 3 are estimators of . We know that E 1 E 2 and , 2 E 3 , Var 1 12, Var 2 10 and E 3 6 Compare these three estimators. Which do you prefer? Why? IE - 2333 SWN 10 4. In a Binomial experiment exactly x successes are observed in n independent trials. The following two statistics are proposed as estimators of the proportion parameter p : T1 X n and T2 X 1 n 2 Determine and compare the MSE for T1 and T2 5. Let X1, X2, X3 and X4 be a random sample of size 4 from a population whose distribution is exponential with unknown parameter . IE - 2333 SWN 11 a. Which of the following statistics are unbiased estimators of ? T1 16 X1 X 2 13 X 3 X 4 T2 T3 X1 2 X 2 3 X 3 4 X 4 5 X1 X 2 X 3 X 4 4 b. Among the unbiased estimators of , determine the one with the smallest variance IE - 2333 SWN 12