Core Concepts Overview
Fundamental behavioral modeling abstractions in BehaviorSim.
BehaviorSim frames behavioral generation around state-machine mechanics. Instead of modeling raw features directly, it models the latent states that produce those features.
Modeling Philosophy
Real systems exhibit behavioral momentum. For example, a user who is confused is likely to remain confused for several interactions before either requesting help or churning. BehaviorSim captures this using Markovian transition matrices and condition rules.
The Three Pillars
1. Discrete States
Non-empty string identifiers representing cognitive, physical, or behavioral phases.
2. Transition Dynamics
Row-stochastic transition probabilities and rule-based history evaluations.
3. Parametric Emissions
Conditioned continuous and discrete distributions emitting observable metrics.
Further Reading
For formal mathematical definitions, visit the public Concepts Page and Methodology Page.