Demographics
Geography France
Gender
Age 38
Financial Profile
Credit Score 650
Balance € 75,000
Estimated Salary € 80,000
Account Details
Tenure (years) 5
Number of Products 2
Has Credit Card
Active Member
Churn Probability
—
MLP output (raw)
Decision
—
Threshold ≥ 0.60
F1 Score
61.9%
Model benchmark
Recall
75.3%
True churn caught
Churn Risk Gauge
Enter inputs and run prediction
Risk Factor Analysis
Factors appear after prediction
Model Architecture
MLP with Dropout regularization layers.
Trained on 10,000 bank customer records. Categorical features encoded via scikit-learn pipeline.
Threshold Tuning
Decision threshold set to 0.60 (vs default 0.5) to optimize
F1-Score — balancing precision and recall for imbalanced churn data.
Key Churn Drivers
Age > 45 and
Geography = Germany are strongest predictors.
Inactive members with high balance are highest risk segment.