Algorithm for model-free feature-free identification of similarity between data streams. Computes a universal, causal similarity distance between hidden stochastic generators from sample paths.

iGillespie is an inverse gillespie algorithm, which infers stochastic reaction systems from data. Gillespie stochastic simulation is used extensively to investigate stochastic phenomena in many fields, ranging from chemistry to biology to ecology. iGillespie solves the inverse problem: How to reconstruct the underlying reactions de novo from sparse intermittent aggregate population counts.

genESeSS constructs propbabilistic graphical models from quantized data streams. The probabilistic automata inferred are not restricted to any specific structural class, and it PAC-efficiently learns any ergodic stationary quantized stochastic process with finite number of causal states.

Algorithm for Estimation of Shannon Entropy of short data-streams. A fundamentally different approach to entropy estimation to that provided by the LZ methods. It works with significantly less data, and more importantly, offers confidence bounds on the computed result.

Paper published in Royal Society Interface:

Article Usage Statistics

Paper published in SIAM Journal of Control \& Optimization: