Stochastic Profiler and Information-Theoretic Entropy Engine for large-scale IEEE OCEL 2.0 event streams.
Computes:
- Shannon Entropy H(L) over trace variants: H(L) = -sum(p_i * log2(p_i))
- Stochastic Transition Probability Matrices P_ij = |(a_i, a_j)| / sum_k |(a_i, a_k)|
- Trace Variant Pareto Cumulative Distribution (top-K variants vs. coverage %)
- Log-normal parameter estimation (mu, sigma) and variance for duration distributions.
Summary
Functions
Streams and computes complete stochastic profile over an NDJSON event log.