Medical Data Analytics Simulator

Patient factors

Age 55 yr

Systolic BP 130 mmHg

Cholesterol 200 mg/dL

10-yr CVD risk –%

Classification

Decision threshold 0.30

Sensitivity

–

Specificity

–

AUC

–

High-risk %

–

How this model works

A simplified logistic-regression cardiovascular-risk model combines standardized patient factors into a linear score z, then squashes it through the sigmoid function to a probability:

z = b0 + w_age·age + w_bp·bp + w_chol·chol + w_smoke·smoke
p = 1 / (1 + e^-z)

Coefficients are illustrative, tuned so risk rises smoothly with age, blood pressure, cholesterol and smoking — the same directional relationships used in real clinical risk scores (e.g. Framingham-style models).
The ROC curve is computed by sweeping the classification threshold across the simulated population and plotting true-positive rate against false-positive rate; AUC summarises overall discrimination.
The population histogram re-simulates 1,000 synthetic patients (factors drawn from randomised distributions) and buckets their computed risk scores, so you can see how the current threshold splits the population into low- and high-risk groups.

Educational demonstration only — not a validated clinical instrument.

Risk curve — p(risk) vs. linear score

ROC curve — sensitivity vs. 1−specificity

Population risk distribution (n = 1000)