Section 48 Multiple Linear Regression Model: Exercise


48.1 Exercise 1:

Blood pressure of individuals: BP.csv

  • Download the Microsoft Excel data file: BP.csv

  • Fit a multiple linear regression model of SBP with all the predictors. Explore all the predictors and possible interaction terms. Use the lm function to fit the model and discuss the anova and summary outputs. Also check the diagnostic plots for model assumptions.

  • Use the matrix formula discussed earlier to develop your own my.lm function to fit a multiple regression model. The function will take the x matrix and y vectors through its arguments, fit a multiple regression model and provide anova and summary outputs.


48.2 Exercise 2:

Spending on different forms of advertisments and sales: Advertising.csv

  • Download the csv data file: Advertising.csv

  • Data show the spending on different forms of advertisments and sales for 200 companies.

  • Develop a multiple linear regression model to assess how different forms of advertisements contributed to sales.

  • Explore model selection and possible interaction between predictors.

  • Use the lm function to fit the model and discuss the anova and summary outputs.

  • Also check the diagnostic plots for model assumptions. Use the model to predict a future sale scenario conditional on mean values of the predictors.


48.3 Exercise 3:

Credit card balance of individuals: Credit.csv

  • Download the csv data file: Credit.csv

  • Data include the credit card balance of 400 individuals and their personal profiles.

  • Develop a multiple linear regression model to identify important predictors that explain the variability in the credit balance of individuals.

  • Use appropriate model section strategies to find the optimal model.

  • Explore the possible interaction between the predictors.

  • Also check the diagnostic plots to validate the model assumptions of the final model.