关键词 > 625.661
625.661 Statistical Models and Regression Modules 1,2 Assignment
发布时间:2023-08-31
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
625.661 Statistical Models and Regression
Modules 1,2 Assignment
Please do all the problems below.
State necessary assumptions for your analytic derivations and analyses.
Read Mod01B. Consider a response variable y and a regressor or
independent variable x . Regardless of whether x is a fixed (i.e., non-
random) variable or a random variable, we can always work on the conditional expectation E( y | x) and conditional variance V( y | x) . An example is a
simple linear regression model
E( y | x) = β + βଵ x
V( y | x) = σଶ ,
where β , βଵ , σ ଶ are all parameters whose values are unknown and do not depend on x . They need to be estimated once we have a random sample.
1. In a simple linear regression analysis of n independent paired data ൫yଵ,xଵ ൯, … . , ൫y,x
൯ to fit the following model labeled M1
E( y | x
) = β
+ βଵ (ax
+ b) ,
ε = y
− E( y
| x
) , i = 1, … , n ,
where the regressor x is a random variable with mean μ௫ and variance σ௫(ଶ) . The random error ε is statistically independent of x and has conditional mean zero and conditional variance σଶ . The terms a and b are real
numbers and a ≠ 0. The values of μ௫ , σ௫(ଶ), σ ଶ are all unknown. Before the n
independent paired data for (y , x) are available, we need to construct estimators for the parameters.
a) Construct the ordinary least squares (OLS) estimator of β1 . Does this estimator change as the value a and/or the value of b are changed?
Provide mathematical proof for your answer.
b) Is the OLS estimator you obtained in a) unbiased for β1 ? Provide mathematical proof for your answer.
c) Construct the ordinary least squares (OLS) estimator of β0 . Does this estimator change as the value a and/or the value of b are changed?
Provide mathematical proof for your answer.
d) Derive the conditional variance of the OLS estimator of β1 in a), i.e., conditional on x’s, and construct an unbiased estimator of this
conditional variance. Provide mathematical proof.
e) Derive the conditional covariance between the OLS estimator of β1 and the OLS estimator of β0 (i.e., conditional on x’s).
f) Find a and b such that conditional on x’s, the correlation between the OLS estimator of β1 and the OLS estimator of β0 is zero.
g) Construct an unbiased estimator for σ2 and provide mathematical proof for unbiasedness.
2. Refer to the linear regression model M1 in Problem 1 above, but x is a non-random regressor. Discuss whether the ordinary least-squares
estimator of the slope β1 is always unbiased and whether it always has the smallest variance compared to any estimator of β1 , irrespectively of what the value of β0 is. State assumptions in your discussion. Be careful about the word “any”.
3. Use any math/stat software (e.g., www.numbergenerator.org/randomnumbergenerator) of your choice to find a random number generator to randomly select 20 rows of Table B.3 for Problem 2.4 of Textbook. Do (a), (b), (c), (d), (e), (f) by hand
calculation (i.e., do not use any software to produce results, but you can use any software to generate the percentile of normal,
t, F, chi-square distribution and/or help with basic mathematical calculations such as add, subtract, divide, multiply, square root).
List which rows of Table B.3 data you selected and provide intermediate steps in your calculations. State assumptions for all steps in your analyses.