Decision trees use a classification or partition as the basis for a predictive model to explore patterns, uncover relationships, screen the most influential factors and make conclusions about a response.
Situation: As the research manager, you want to understand if there is a relationship between client factors and credit card offers.
Task: The problem is that there are multiple customer factors and not all the relationships, distributions and interactions are fully understood.
Action: You use Decision Trees to create a simple Partition Analysis then move onto advanced Bootstrap Forest and Boosted Trees methods to find the best model.
Result: Using these Machine Learning techniques you compare the best models and identify the optimal values to maximize acceptance.
Next Steps: Your research is shared interactively with your colleagues and informs the next marketing campaign and improves profitability.
https://www.jmp.com/en_au/academic/case-study-library.html#credit
