CNF Rear Geometry
Vicente Cohort Fit Engine v3
Geometry-aware cohort similarity, style scoring, and fit recommendation system anchored to the Vicente centroid.
Cohort Definition
Users whose metrics fall within the Vicente centroid:
- Projection: 22–34 mm
- Curvature: 0.80–1.05
- Contour stability: ≥ 0.85
- FitBand: regular or relaxed
- Tension hotspots: ≤ 3 zones
Fit Engine Input Vector
x = [μ_κ, σ_κ, μ_p, σ_p, FitBand, T_zones, preferences]
Where μ_κ = mean curvature, σ_κ = curvature std, μ_p = mean projection, σ_p = projection std, T_zones = tension per zone.
Cohort Similarity
d(x, c_Vicente) = ||x − c_Vicente||₂
Lower distance indicates stronger cohort match. Used to rank Levi's cuts by geometry compatibility.
Style Scoring
score(s, x) = w_g · geom_match + w_c · cohort_success + w_p · pref_match
The engine returns top Levi's styles with size adjustments, comfort predictions, and engineering notes.
Why It Matters
- Geometry-aware Levi's recommendations replace generic waist sizing.
- Tension-aware size suggestions reduce return rates.
- Comfort-optimized cut selection drives cohort satisfaction.
- Pattern feedback loops close the gap between lab and store.