Deck 12: Regression Analysis II Multiple Linear Regression and Other Topics

Question 1

Find a linearizing transformation of y=1/(a + b x)

Answer 1

y' = 1/y, x' = x

Question 2

Consider the data set

A) Obtain the best fitting straight line with
B) What proportion of the y'variability is explained by the fitted line?
Round your answers to three decimal places.

Answer 2

Part A: = .180 - 1.940x
Part B: .993

Question 3

Suppose the price, y, per ton of copper can be modeled by where the two predictor variables x₁ and x₂ are the prices per ton of zinc and lead, respectively. The least squares estimates are:
Predict the response for x₁= 2389 and x₂ = 2366.

Answer 3

$7131.07

Question 4

Consider the multiple linear regression model

where , and the normal random variable e has standard deviation 6. What is the mean of the response Y when x₁ = 1 and x₂ = 4?

A) 34
B) 19
C) 40
D) 25

Question 5

Suppose the price, y, per ton of copper can be modeled by
where the two predictor variables x₁ and x₂ are the prices per ton of zinc and lead, respectively. The least squares estimates, based on monthly observations in a five-year period, are:
Assuming that the residual sum of squares (SSE) is 1424.48 and the SS due to regression is 5150.68.

A) Estimate the error standard deviation.
B) State the degrees of freedom used in part A.
C) Find R².