A First-Principles Hybrid Attribution Framework

A First-Principles Hybrid Attribution Framework Integrating Probabilistic Path Modeling, Cooperative Game Theory, and Psychographic Transition Priors Traditional attribution models fail by either ignoring sequence (heuristic methods) or sacrificing fairness (pure probabilistic models). This paper presents a hybrid framework that combines absorbing Markov chains for causal sequence modeling with Shapley Value attribution for axiomatic fairness, augmented by psychographic transition priors for behavioral context. The framework is device-agnostic, identifier-independent, and grounded in first principles of stochastic processes and cooperative game theory. We provide a rigorous theoretical foundation with explicit assumptions, a frozen reference implementation (v1.0.0), comprehensive sensitivity and uncertainty quantification, and clear guidance on when attribution results are trustworthy.