precise methods relate to neighboring research —
drawn as graphs, with the missing connections marked explicitlyEach map is an interactive graph: nodes are papers, colors are literatures, solid edges are established connections, and dashed red edges are verified absences — combinations no one has published, i.e. research directions.
The damped conditional Gaussian factorization at the center; Cotton 2022/2024 and Chakraborty–Katzfuss (ShrinkTM) in the core circle; a ring of derived works citing them; and ten literatures fanning out — Vecchia approximations, composite likelihood, Cholesky regressions, shrinkage and random matrix theory, robust covariance, positive-definite completion, portfolio construction, machine learning, and scoring rules. Also rendered on schur.microprediction.org, with a companion timeline.
Planned: robust covariance at massive scale (heavy tails, cellwise contamination, streaming); covariance evaluation and scoring rules (which judge for which gold). Maps are hand-curated; the inclusion rule is a verified methodological connection or a formal citation — file an issue on microprediction/precise to propose an addition.