Example of Multivariate Prediction

  1. The examples on multivariate prediction are meant to help you become familiar with the concepts, data sets, use of R, etc.
  2. Create folder MULTIVARIATE with data sets in it, so that the directory of R can be set to access this folder and the files in it.
  3. Please use Wordpad to open the dataset, if needed.
  4. The Answer 3.1, Answer 3.5 and Answer 3.7 files are example answers by students, to help you follow the line by line code development.
  5. Use the Rprimer for additional help
  6. Please let me know if you have any difficulties in working on these assignments.
    ------
  1. Page 78, Exercise 3.1, Analyzing Multivariate Data, Lattin, Carroll, Green (PCG)

The data on heights and weights of women used in the chapter used in the chapter to illustrate the intuition underlying the simple regression model are available in the file WOMEN. Using these data, test the hypothesis that there is no relationship between a woman’s height and weight.

See file ANSWER 3.1 for the steps of R.

  1. Page 78-79, LCG, Exercise 3.5, LCG.

The file RADIOLOGY contains information on the total number of radiology visits at a particular hospital over a 31-month period. The variables in the data set are defined in Table 3.15 below:

Table 3.15: Variables in Radiology data set

Variable Definition

RAD_VIS Number of radiology visits during the month

P_DAYS Number of patient days during the month

ER_VIS Number of emergency room visits during the month

CL_VIS Number of clinic visits during the month

  1. Plot the relationship between the radiology visits and each of the other variables in the data set. Which of the three variables (P_DAYS, ER_VIS, or CL_VIS) seems to be the strongest determinant of radiology visits? How would you describe the relationship?
  2. Regress radiology visits on the three remaining variables in the data set. How do your conclusions change (if at all) regarding the relationship between the dependent variable an each independent variable?