◈ CV RISK CLUSTER EXPLORER
10 cardiovascular disease models · visit-based score simulation · 6 clustering algorithms
① Data & Prevalence
② Clustering
③ Positivity Analysis
Disease Prevalence & Model Parameters
N patients
Seed
Sensitivity
0.85
Specificity
0.90
▶ GENERATE DATASET
Score Distributions (positive vs negative class)
ALGORITHM
K-Means
GMM — Gaussian Mixture
Fuzzy C-Means
Spectral Clustering
Archetypal Analysis ★
Self-Organizing Map (SOM)
▶ CLUSTER
PCA 2D Projection — colored by cluster
Cluster Profiles — mean model score per disease
Generate a dataset and run clustering to see results.
Classification Thresholds (score cutoff per disease)
Positivity Heatmap — % patients above threshold per cluster × disease
Cluster Sizes