arXiv subjects

Shariq Mohammed

Publications and source records attributed to Shariq Mohammed.

8 recordsLinked to original sources

arXiv2026

Bayesian Feature Extraction using Gaussian and Diffused-gamma Priors for High Dimensional Spatio-Temporal Data

High-dimensional data with sparse structure and spatio-temporal dependence arise in many scientific domains. We develop a Bayesian feature-extraction framework for spatio-temporal settings that employs Gaussian and Diffused-gamma priors to induce structured sparsity. The modeling framework specifies a general likelihood via Bregman divergence, enabling compatibility with a range of loss functions and measurement models. Posterior computation is carried out via Markov chain Monte Carlo (MCMC), and we introduce a two-stage feature-extraction procedure based on posterior samples to stabilize selection across space and time. We illustrate the method with a multi-subject electroencephalography (EEG) case study examining the relationship between chronic alcohol exposure and activity in different brain regions. We first fit binary classification models at each time point, then use false discovery rate-controlled screening and subsequent clustering in a two-stage feature-extraction pipeline to identify active brain regions. The analysis demonstrates that our proposed priors improve recovery of sparse features and enhance interpretability in the presence of spatio-temporal dependence. The framework is broadly applicable to high-dimensional, structured problems where accurate feature selection and inference are required. The code to implement the model is publicly available via GitHub.

stat.ME↗

arXiv2024

Statistical Analysis of Quantitative Cancer Imaging Data

Recent advances in types and extent of medical imaging technologies has led to proliferation of multimodal quantitative imaging data in cancer. Quantitative medical imaging data refer to numerical representations derived from medical imaging technologies, such as radiology and pathology imaging, that can be used to assess and quantify characteristics of diseases, especially cancer. The use of such data in both clinical and research setting enables precise quantifications and analyses of tumor characteristics that can facilitate objective evaluation of disease progression, response to therapy, and prognosis. The scale and size of these imaging biomarkers is vast and presents several analytical and computational challenges that range from high-dimensionality to complex structural correlation patterns. In this review article, we summarize some state-of-the-art statistical methods developed for quantitative medical imaging data ranging from topological, functional and shape data analyses to spatial process models. We delve into common imaging biomarkers with a focus on radiology and pathology imaging in cancer, address the analytical questions and challenges they present, and highlight the innovative statistical and machine learning models that have been developed to answer relevant scientific and clinical questions. We also outline some emerging and open problems in this area for future explorations.

stat.AP↗

arXiv2023

Bayesian Variable Selection in Double Generalized Linear Tweedie Spatial Process Models

Double generalized linear models provide a flexible framework for modeling data by allowing the mean and the dispersion to vary across observations. Common members of the exponential dispersion family including the Gaussian, Poisson, compound Poisson-gamma (CP-g), Gamma and inverse-Gaussian are known to admit such models. The lack of their use can be attributed to ambiguities that exist in model specification under a large number of covariates and complications that arise when data display complex spatial dependence. In this work we consider a hierarchical specification for the CP-g model with a spatial random effect. The spatial effect is targeted at performing uncertainty quantification by modeling dependence within the data arising from location based indexing of the response. We focus on a Gaussian process specification for the spatial effect. Simultaneously, we tackle the problem of model specification for such models using Bayesian variable selection. It is effected through a continuous spike and slab prior on the model parameters, specifically the fixed effects. The novelty of our contribution lies in the Bayesian frameworks developed for such models. We perform various synthetic experiments to showcase the accuracy of our frameworks. They are then applied to analyze automobile insurance premiums in Connecticut, for the year of 2008.

stat.ME↗

arXiv2022

Integrative Bayesian models using Post-selective Inference: a case study in Radiogenomics

Integrative analyses based on statistically relevant associations between genomics and a wealth of intermediary phenotypes (such as imaging) provide vital insights into their clinical relevance in terms of the disease mechanisms. Estimates for uncertainty in the resulting integrative models are however unreliable unless inference accounts for the selection of these associations with accuracy. In this article, we develop selection-aware Bayesian methods which: (i) counteract the impact of model selection bias through a "selection-aware posterior" in a flexible class of integrative Bayesian models post a selection of promising variables via $\ell_1$-regularized algorithms; (ii) strike an inevitable tradeoff between the quality of model selection and inferential power when the same dataset is used for both selection and uncertainty estimation. Central to our methodological development, a carefully constructed conditional likelihood function deployed with a reparameterization mapping provides notably tractable updates when gradient-based MCMC sampling is used for estimating uncertainties from the selection-aware posterior. Applying our methods to a radiogenomic analysis, we successfully recover several important gene pathways and estimate uncertainties for their associations with patient survival times.

stat.AP↗

arXiv2021

Tumor Radiogenomics with Bayesian Layered Variable Selection

We propose a statistical framework to integrate radiological magnetic resonance imaging (MRI) and genomic data to identify the underlying radiogenomic associations in lower grade gliomas (LGG). We devise a novel imaging phenotype by dividing the tumor region into concentric spherical layers that mimics the tumor evolution process. MRI data within each layer is represented by voxel--intensity-based probability density functions which capture the complete information about tumor heterogeneity. Under a Riemannian-geometric framework these densities are mapped to a vector of principal component scores which act as imaging phenotypes. Subsequently, we build Bayesian variable selection models for each layer with the imaging phenotypes as the response and the genomic markers as predictors. Our novel hierarchical prior formulation incorporates the interior-to-exterior structure of the layers, and the correlation between the genomic markers. We employ a computationally-efficient Expectation--Maximization-based strategy for estimation. Simulation studies demonstrate the superior performance of our approach compared to other approaches. With a focus on the cancer driver genes in LGG, we discuss some biologically relevant findings. Genes implicated with survival and oncogenesis are identified as being associated with the spherical layers, which could potentially serve as early-stage diagnostic markers for disease monitoring, prior to routine invasive approaches.

stat.ME↗

arXiv2021

RADIOHEAD: Radiogenomic Analysis Incorporating Tumor Heterogeneity in Imaging Through Densities

Recent technological advancements have enabled detailed investigation of associations between the molecular architecture and tumor heterogeneity, through multi-source integration of radiological imaging and genomic (radiogenomic) data. In this paper, we integrate and harness radiogenomic data in patients with lower grade gliomas (LGG), a type of brain cancer, in order to develop a regression framework called RADIOHEAD (RADIOgenomic analysis incorporating tumor HEterogeneity in imAging through Densities) to identify radiogenomic associations. Imaging data is represented through voxel intensity probability density functions of tumor sub-regions obtained from multimodal magnetic resonance imaging, and genomic data through molecular signatures in the form of pathway enrichment scores corresponding to their gene expression profiles. Employing a Riemannian-geometric framework for principal component analysis on the set of probability densities functions, we map each probability density to a vector of principal component scores, which are then included as predictors in a Bayesian regression model with the pathway enrichment scores as the response. Variable selection compatible with the grouping structure amongst the predictors induced through the tumor sub-regions is carried out under a group spike-and-slab prior. A Bayesian false discovery rate mechanism is then used to infer significant associations based on the posterior distribution of the regression coefficients. Our analyses reveal several pathways relevant to LGG etiology (such as synaptic transmission, nerve impulse and neurotransmitter pathways), to have significant associations with the corresponding imaging-based predictors.

stat.ME↗

arXiv2020

Spatial Tweedie exponential dispersion models

This paper proposes a general modeling framework that allows for uncertainty quantification at the individual covariate level and spatial referencing, operating withing a double generalized linear model (DGLM). DGLMs provide a general modeling framework allowing dispersion to depend in a link-linear fashion on chosen covariates. We focus on working with Tweedie exponential dispersion models while considering DGLMs, the reason being their recent wide-spread use for modeling mixed response types. Adopting a regularization based approach, we suggest a class of flexible convex penalties derived from an un-directed graph that facilitates estimation of the unobserved spatial effect. Developments are concisely showcased by proposing a co-ordinate descent algorithm that jointly explains variation from covariates in mean and dispersion through estimation of respective model coefficients while estimating the unobserved spatial effect. Simulations performed show that proposed approach is superior to competitors like the ridge and un-penalized versions. Finally, a real data application is considered while modeling insurance losses arising from automobile collisions in the state of Connecticut, USA for the year 2008.

stat.ME↗

arXiv2020

Spatial risk estimation in Tweedie compound Poisson double generalized linear models

Tweedie exponential dispersion family constitutes a fairly rich sub-class of the celebrated exponential family. In particular, a member, compound Poisson gamma (CP-g) model has seen extensive use over the past decade for modeling mixed response featuring exact zeros with a continuous response from a gamma distribution. This paper proposes a framework to perform residual analysis on CP-g double generalized linear models for spatial uncertainty quantification. Approximations are introduced to proposed framework making the procedure scalable, without compromise in accuracy of estimation and model complexity; accompanied by sensitivity analysis to model mis-specification. Proposed framework is applied to modeling spatial uncertainty in insurance loss costs arising from automobile collision coverage. Scalability is demonstrated by choosing sizable spatial reference domains comprised of groups of states within the United States of America.

stat.AP↗