Question

Please explain the following.

1. A transformation can be used to:

(a) account for curvature

(b) get a better prediction equation

(c) stabilize the variances

(d) any of the above

2. A partial regression plot that shows a straight-line relationship indicates:

(a) that a linear term needs to be added to the model

(b) that a quadratic term needs to be added to the model

(c) the linear relationship that already exists in the model

(d) all of the above

3. The F test statistic from a lack of fit test should equal

(a) the overall model F test statistic

(b) the square of the t test statistic for the intercept

(c) 1/(1-R²)

(d) none of the above

4. If in a regression model we have accounted for every source of variability except random error, then it must be true that:

(a) R²=1

(b) SS_LOF =0

(c) Both (a) and (b)

(d) Neither (a) nor (b)

5. A cross validation data set should be

(a) A subset of the data set used to build the model

(b) Similar to the data set used to build the model

(c) Smaller than the data set used to build the model

(d) None of the above

Answer 1