Your Navigation Path
Case Base

Case: Churn Prediction Case

Has Chains Part 1 - Treat missing values in CDR
Part 2 - Transpose CDR from transactional to relational form
Part 3 - Transpose REVENUES from transactional to relational form
Part 4 - Create derived attributes and customer profile
Part 5 - Churn Modeling

M4 Documentation:

A more complex Case in which decision trees are learned to predict the churn (contract dissolution) behaviour of telecommunication clients. Input data are call detail records (CDR). Because the case was created with an earlier version of the MiningMart system, where the operator Pivotize was not available, it uses a rather large number of other operators to implement pivotisation and thus demonstrates the usefulness of Pivotize. Note also that for the same reason, the concept links are not set in this case, so that the concept icons appear unrelated in the concept editor.

Additional documentation for the Case Base: