Dynamics

Dynamics

Methods for modeling change over time and space from repeated measurements.

This site reviews differential-equation, supervised-learning, and hybrid approaches to longitudinal modeling along two design axes: the number of observations per subject, and the number of spatially structured variables measured at each one. It states which models are identifiable in a given design, which require pooling across subjects, and which are not estimable at any sample size — and it treats spatial structure as something to be established from the data rather than imposed from prior knowledge.

E.D. Gennatas MBBS AICSM PhD + AI
rtemis.org