Course Evaluation:

Please take a moment to fill out Monmouth's Course Evaluation for this class. Your feedback will help me to improve the course in future terms.

Multiple Linear Regression and Model Selection:

Important: You will need to use Minitab on a PC, not Minitab Express on a Mac, to complete this assignment. Minitab Express, as the name suggests, does not have all the functionality that of Minitab.

Download this Minitab project containing the data we used for body fat prediction in the previous three class sessions.

The goal of this assignment is to determine the best model for predicting body fat percentage from various other biometrics using a multiple linear regression. The response is body fat percentage, and we will consider the following predictors:

• age
• weight
• height
• BMI
• neck
• chest
• abdomen
• hip
• thigh
• knee
• bicep
• ankle
• forearm
• wrist

Assume that all measurements after BMI are circumferences of the body part in question.

1. Generate a Matrix Plot:
   1. Use Minitab to to generate matrix plot showing each of the predictors plotted against themselves and against the response.
      • From the top menu bar, select Stat > Matrix Plot....
      • Select the Simple option from the second row.
      • In the Y variables: text box, include the response.
      • In the X variables: text box, enter the predictors listed above
   2. Print out the matrix plot.
2. Fit a Multiple Linear Regression with All of the Predictors:
   1. Use Minitab to estimate the multiple linear regression model that includes all of the candidate predictors at once, using Stat > Regression > Regression > Fit Regression Model....
      • Important: Be sure to also generate the diagnostic plots by clicking the Graphs... button in the dialog box, and selecting the Four in one radio button from the resulting window.
   2. Print the session window and the diagnostic plots.
   3. Annotate the printout with the estimate of the linear regression model.
   4. Annotate the printout with your interpretation of the coefficients and intercept of the linear regression model.
   5. Give your interpretation of the diagnostic plots. Does it appear that the assumptions about how the residuals should behave to make inferences hold for this data? If not, how do the diagnostic plots show deviation from the violated assumption(s)?