Scholars@Duke
Omar Melikechi

Omar Melikechi

Assistant Professor of Statistical Science
Statistical Science

Scholarly Works

Local graph estimation with pathwise false discovery control.

Journal article Nature communications · May 2026 Many datasets include a small set of variables, such as biomarkers or clinical outcomes, whose relationships to the broader system are of primary scientific interest. Estimating the full network of inter-variable relationships in such settings often obscur ... Full text Cite

Sequential Gibbs posteriors with applications to principal component analysis.

Journal article Biometrika · January 2026 Gibbs posteriors are proportional to a prior distribution multiplied by an exponentiated loss function, with a key tuning parameter that weights the information in the loss relative to the prior and provides control of posterior uncertainty. Gibbs posterio ... Full text Cite

Integrated path stability selection.

Journal article Journal of the American Statistical Association · September 2025 Stability selection is a popular method for improving feature selection algorithms. One of its key attributes is that it provides theoretical upper bounds on the expected number of false positives, E(FP), enabling false positive control in practice. Howeve ... Full text Open Access Cite

Nonparametric IPSS: fast, flexible feature selection with false discovery control.

Journal article Bioinformatics (Oxford, England) · May 2025 MotivationFeature selection is a critical task in machine learning and statistics. However, existing feature selection methods either (i) rely on parametric methods such as linear or generalized linear models, (ii) lack theoretical false discovery ... Full text Open Access Cite

Ellipsoid fitting with the Cayley transform.

Journal article IEEE transactions on signal processing : a publication of the IEEE Signal Processing Society · January 2024 We introduce Cayley transform ellipsoid fitting (CTEF), an algorithm that uses the Cayley transform to fit ellipsoids to noisy data in any dimension. Unlike many ellipsoid fitting methods, CTEF is ellipsoid specific, meaning it always returns elliptic solu ... Full text Open Access Cite

Random Splitting of Fluid Models: Unique Ergodicity and Convergence

Preprint · July 1, 2023 We introduce a family of stochastic models motivated by the study of nonequilibrium steady states of fluid equations. These models decompose the deterministic dynamics of interest into fundamental building blocks, i.e., minimal vector fields preserving som ... Full text Open Access Cite

Hitting time of Brownian motion subject to shear flow

Journal article Involve · January 1, 2022 The 2-dimensional motion of a particle subject to Brownian motion and ambient shear flow transportation is considered. Numerical experiments are carried out to explore the relation between the shear strength, box size, and the particle’s expected first hit ... Full text Open Access Cite