Dhrubajyoti Ghosh

Statistics · Data Science · Biomedical Methodology

Dhrubajyoti Ghosh, PhD

Assistant Professor of Data Science and Analytics at Kennesaw State University. I develop statistical and machine-learning methods for longitudinal studies, clinical trials, causal inference, nonlinear time series, and complex biomedical data.

Research Publications LINDT Lab Download CV

Google Scholar · GitHub · LinkedIn · ORCID

Research program

Statistical methodology for complex longitudinal and causal data

My work sits at the intersection of rigorous statistical inference, modern computation, and biomedical applications. A recurring goal is to develop methods that remain reliable when data are multivariate, longitudinal, high-dimensional, non-Gaussian, network-structured, or incompletely observed.

L

Longitudinal & Clinical Trials

Rank-based inference, multivariate longitudinal outcomes, missing data, trial design, power, and covariate adjustment.

I

Imaging & Neurodegeneration

Alzheimer’s, Parkinson’s, MRI/CT, survival analysis, digital twins, and virtual trials.

N

Nonparametric & Robust Inference

U-statistics, rank methods, global testing, resampling, and robust inference for complex data.

D

Data Science, Causal Learning & AI

Causal discovery, IV/MR, graph learning, AI/misinformation, and structured high-dimensional systems.

T

Time Series & Higher-Order Dependence

Polyspectra, nonlinear processes, quadratic prediction, and higher-order frequency-domain inference.

Selected work

Representative publications

Polyspectral Mean Estimation of General Nonlinear Processes
Bernoulli (2026) · D. Ghosh, T. McElroy, S. Lahiri
Higher-order spectral inference for general nonlinear processes.
Penalized FCI for Causal Structure Learning in a Sparse DAG for Biomarker Discovery in Parkinson’s Disease
Annals of Applied Statistics (accepted/in press) · S. Pal, D. Ghosh, S. Yang
Preprint · R package
A non-parametric U-statistic testing approach for multi-arm clinical trials with multivariate longitudinal data
Journal of Multivariate Analysis 209, 105447 (2025) · D. Ghosh, S. Luo
Article · Software
Power and sample size calculation for multivariate longitudinal trials using the longitudinal rank sum test
Statistics in Medicine 44, e70261 (2025) · D. Ghosh, X. Xu, S. Luo
PubMed · Software
Ensemble survival analysis for preclinical cognitive decline prediction in Alzheimer’s disease using longitudinal biomarkers
Journal of Alzheimer’s Disease (2025) · D. Ghosh, S. Pal, M. Lutz, S. Luo, ADNI
Biomedical application of longitudinal biomarkers and survival learning.

View all publications →

Research group

LINDT Lab

LINDT stands for Longitudinal, Imaging, Nonparametric, Data Science & Time-Series Lab. The group develops statistical and computational methods for extracting reliable evidence from complex scientific data, with emphasis on longitudinal inference, causal structure, biomedical applications, and modern data science.

Meet the lab →

Current directions

What we are working on

Robust longitudinal inference

Extending rank-based longitudinal methods to covariate adjustment, incomplete data, observational studies, and complex trial designs.

Network-aware causal learning

Combining instrumental variables, graph regularization, and latent/network structure for high-dimensional causal discovery and estimation.

Biomedical data science

Methods and applications in Alzheimer’s and Parkinson’s disease, medical imaging, digital twins, survival prediction, and clinical-trial analysis.

Nonlinear & structured data

Polyspectral inference, nonlinear dependence, spatial/environmental health, graph learning, and AI-assisted statistical workflows.

Contact

Dhrubajyoti Ghosh
Assistant Professor, School of Data Science and Analytics
Kennesaw State University
dghosh3@kennesaw.edu