14.4 MINING ASSOCIATION RULES
So, let us get our hands dirty mining for association rules using the Churn_Training_File data set. Prepare by doing the following:
- Subset the following variables into their own data frame: VMail Plan, Intl Plan, CustServ Calls, and Churn.
- Set CustServ Calls to be an ordinal factor.
Let us begin by finding the “baseline” proportions for the various variables, so that we may later check the confidence levels of our association rules against these baseline levels. These proportions may be found in Figures 14.1 and 14.2. For example, the proportion of customers who churn is 14.53%.
Now, let us generate some association rules, using the following settings:
- Specify the type of association to obtain as “rules”
- Minimum support equals 0.01 (1%)
- Minimum confidence equals 0.4 (40%)
- Maximum...