Use

>install.packages("faraway")

>library(faraway)

>data(wbca)

in r to access the data.

Problems 4 and 5 will use Faraway's wbca data on diagnosis of breast cancer among women with breast tumors using measures of the abnormality of cell characteristics. Study the help file carefully to gain some understanding of the variables.

4. Use logistic regression to model the probability a tumor is malignant (cancerous) using all the variables except Class as predictors. You might want to recode the data to make this easier using a response variable malignant=1*(Class==0). a. Find the probability a tumor is malignant (cancerous) when each predictor takes the value 3. b. Find a 95 percent confidence interval for the probability estimated in (a).

5. Decide you will classify a tumor as malignant if the fitted probability for being malignant is 0.5 or greater. a. Estimate the sensitivity of this test. b. Estimate the specificity of this test.