Retail Bank Customer Churn

Client Model

Retail banking machine learning model designed to identify clients that are likely to “churn” and migrate their day to day banking to competitors

Client Model Prediction

Predicted ClassChurn
Model Confidence93%
R-Squared0.64
Significance0.0001

DC-MINT – AI EXPLAINABILITY

  • The customer was classed as “Churn”
  • The top three contributing factors towards the ‘Churn’ prediction were – Gender, Age and Number of products
  • To estimate the impact that changing each driver of churn would have in the probability of churning for retail bank
  • To enable retention teams to focus on these drivers and tailor responses for retail bank

DC-MINT INsight