Glossary
Archetypal Clustering
Unsupervised learning to identify behavioral clusters like 'The Skeptical Innovators' rather than demographic groups.
Definition
Archetypal Clustering is an unsupervised machine learning approach that discovers behavioral segments-archetypes-from data without predefined labels. Instead of demographic segments like 'Males 30-40,' archetypal clustering reveals behavioral clusters: 'The Skeptical Innovators,' 'The Impulse Explorers,' segments that predict action. Sentient OS applies this in Layer 5 (Pattern Recognition) of the architecture. The algorithm operates in vector space: persona vectors cluster by similarity, and clusters are interpreted as archetypes. These archetypes drive audience intelligence, creator matching, and content resonance modeling. Archetypal clustering transcends demographics because behavior-not age or gender-predicts conversion, engagement, and fit. No labels required; the data reveals the segments.
Why It Matters
Archetypal clustering is how Sentient discovers 'The Skeptical Innovators.' It's the foundation of audience intelligence that goes beyond demographics.
Related Pages
Related Terms
Unsupervised Learning
An AI approach that discovers hidden patterns without needing labeled examples. Used to find behavioral groups that demographics miss. For behavioral archetype identification.
Persona Vectors
Mathematical representations of customers as points in complex space, enabling computable similarity and distance.
Customer Segmentation
Dividing audiences into groups. Sentient goes beyond age and gender to behavioral patterns that actually predict response.
Psychographic Profiling
Understanding attitudes, interests, personality traits. Sentient's Layer 4 creates mathematical psychographic models.
Vector Spaces
A mathematical space where people, products, and content are represented so that 'closeness' means compatibility. The foundation for precise matching. Mathematics instead of databases.
Explore the Full Platform
See how these concepts come to life inside Sentient OS.