PREDICTIVE ANALYTICS ENGINE

Machine learning models for churn prediction, loan propensity, and customer lifetime value

Model Accuracy: 85% prediction accuracy

Validated on historical data with cross-validation

Require immediate attention

$167,188

Potential revenue at risk

$117,032

Revenue that can be saved

90 days

Early warning period

Customer count by risk level

Detailed risk distribution

Low Risk

145 customers

Medium Risk

42 customers

High Risk

13 customers

Highest churn probability scores (next 90 days)

Customer	Segment	Churn Risk	CLV	Risk Level
Jennifer Williams CUST-0162	Credit Seekers	33.2%	$8,352	HIGH
Patricia Brown CUST-0048	Balanced Customers	32.9%	$5,842	HIGH
Michael Smith CUST-0023	Credit Seekers	32.4%	$20,443	HIGH
Patricia Miller CUST-0077	Balanced Customers	31.8%	$6,297	HIGH
James Johnson CUST-0199	High-Value Professionals	31.4%	$24,858	HIGH
James Martinez CUST-0155	Balanced Customers	30.7%	$21,591	HIGH
Elizabeth Smith CUST-0034	Balanced Customers	30.4%	$19,906	HIGH
John Brown CUST-0141	High-Value Professionals	29.1%	$7,282	HIGH
Linda Williams CUST-0053	Young Savers	28.9%	$10,408	HIGH
John Smith CUST-0136	Balanced Customers	27.7%	$14,441	HIGH
Robert Brown CUST-0044	High-Value Professionals	26.8%	$18,253	HIGH
Michael Jones CUST-0075	Balanced Customers	26.1%	$698	HIGH
Robert Smith CUST-0148	Balanced Customers	25.5%	$8,815	HIGH
Jennifer Rodriguez CUST-0180	Balanced Customers	23.4%	$17,833	MEDIUM
Robert Brown CUST-0182	Balanced Customers	23.1%	$22,701	MEDIUM
James Johnson CUST-0128	High-Value Professionals	22.8%	$19,068	MEDIUM
John Williams CUST-0176	Balanced Customers	21.9%	$6,502	MEDIUM
Robert Johnson CUST-0191	Balanced Customers	21.5%	$12,667	MEDIUM
Elizabeth Rodriguez CUST-0146	Balanced Customers	21.3%	$12,533	MEDIUM
Elizabeth Miller CUST-0170	Credit Seekers	20.8%	$16,653	MEDIUM