Scribd
Upload
20 views

HW1 Soln

1. The document discusses calculating marginal and joint probability density functions and determining if two random variables are independent. 2. Expected values, variances, covariances, and correlations are calculated for several random variables to determine if one is a linear function of another. 3. Maximum likelihood and method of moments estimators are calculated and compared for a parameter of a probability distribution. 4. Confidence intervals are derived for a population mean using small and large sample procedures and their differences are noted. A hypothesis test is performed to test if a population mean is equal to a given value.

Uploaded by

tysoccer12
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
20 views

HW1 Soln

1. The document discusses calculating marginal and joint probability density functions and determining if two random variables are independent. 2. Expected values, variances, covariances, and correlations are calculated for several random variables to determine if one is a linear function of another. 3. Maximum likelihood and method of moments estimators are calculated and compared for a parameter of a probability distribution. 4. Confidence intervals are derived for a population mean using small and large sample procedures and their differences are noted. A hypothesis test is performed to test if a population mean is equal to a given value.

Uploaded by

tysoccer12
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Solution to HW1. 1. (i)First nd the equation for the hypotenuse y = 3 0.5x.

. Then notice from the domain of the joint density that trivially FX (x) = 0 for x < 2, FX (x) = 1 for x 4 and FY (y ) = 0 for y < 1, FY (y ) = 1 for y 2. It is only necessary to calculate: x 3 .5 x 1dydt = (x 2)(3 .5x)/2. For 2 x < 4, FX (x) = 2 1 y 6 2s 1dxds = (y 1)(3 y ). For 1 y < 2, FY (y ) = 1 2 (ii) Dierentiate the marginal CDF, fX (x) = FX (x) = 2 .5x for 2 x < 4 and fY (y ) = FY (y ) = 4 2y for 1 y < 2. When we look at the product fX (x)fY (y ), we see that this product depends on the specic (x, y ) values hence not equal to the joint pdf fX,Y . So one can conclude that X and Y are not independent. 2. (i) E (X ) = 0 .23+2 .1+(1) .5+3 .17 = .21, E (Y ) = 1 .23+5 .1+ (1).5+7.17 = 1.42, E (XY ) = 0.23+25.1+(1)(1).5+37.17 = 5.07. (ii) First calculate Var(X ) = E (X 2 ) .212 = 0 .23 + 4 .1 + 1 .5 + 9 .17 .212 = 2.3859, Var(Y ) = E (Y 2 ) 1.422 = 1 .23 + 25 .1 + 1 .5 + 49 .17 1.422 = 9.5436. Then calculate Cov(X, Y ) = E ( XY ) E (X )E (Y ) = 5.07 .21 1.42 = 4.7718 and Corr(X, Y ) = 4.7718/ 2.3859 9.5436 = 1. It means Y is a linear function of X and vice versa. 3. (i) Calculate E (X ) = 0 ( + 1)x=1 dx = ( + 1)/( + 2). Let ( + 1)/( + = (2X and solve for . The MM estimator is 1)/(1 X ) = 3. 2) = X n (ii) The log likelihood function is l(x; ) = ln[( +1) ( Xi ) ]. Let l (x; ) = 0 = (n + ln( Xi ))/ ln( Xi ) = 3.116. and solve for . The MLE is 4. (i) Find that t.025,14 = 2.145. The 95% condence interval for the mean failure strain by a small sample procedure is: 25 2.145 3.5/ 15 = (23.06, 26.94). (ii) The 95% condence interval is:25 2.145 3.5 1 + 1/15 = (17.24, 32.75. It is bigger than the CI given in part (i) because here we are predicting the behavior of a new R.V. instead of estimating the constant population mean. 5. This is a large sample z -test. The test statistic is z = (4.875)/(.35/ 100) = 3.71. The P-value for testing the alternative hypothesis = 5 is 2 P [z <
1

3.71] = 0. We must reject the null hypothesis when using = 0.01 given this very small P-value.

You might also like