Practical Predictive Analytics

Comparing outcomes

As mentioned in an earlier chapter, boxplots are a good way for comparing two numeric variables. We will compare Diabetes='neg' with Diabetes='pos'.

Position your cursor within the next cell and run the following code chunk. This code runs the boxplot() function for variables 1-7, and breaks it out by diabetes outcome:

#some basic exploratory analysis.  Box plots by outcome par(mfrow=c(2,4)) for (i in 1:7) {   boxplot(PimaIndiansDiabetes[,i] ~ PimaIndiansDiabetes$diabetes, main=names(PimaIndiansDiabetes[i]), type="l")    } 

After the run is complete (it should only take a second), the following boxplots will appear under the code that was just run:

In the boxplots produced, you can see that the median values for all of the ...