A5038524961- Complex Network Analysis Techniques
- Bayesian Methods and Mixture Models
- Bayesian Modeling and Causal Inference
- Bioinformatics and Genomic Networks
- Topological and Geometric Data Analysis
- Statistical Methods in Clinical Trials
- Gastric Cancer Management and Outcomes
- Data Management and Algorithms
- Statistical Methods and Bayesian Inference
- Advanced Clustering Algorithms Research
- Pharmacogenetics and Drug Metabolism
- Esophageal Cancer Research and Treatment
- Advanced Causal Inference Techniques
- Statistical Methods and Inference
- Marine and coastal ecosystems
- HIV, Drug Use, Sexual Risk
- Cancer Treatment and Pharmacology
- Data Visualization and Analytics
- Mental Health Research Topics
- Colorectal Cancer Surgical Treatments
- Lymphoma Diagnosis and Treatment
- Colorectal and Anal Carcinomas
- Colorectal Cancer Treatments and Studies
- 3D Shape Modeling and Analysis
- Opioid Use Disorder Treatment
Instituto Tecnológico Autónomo de México
2023
Lancaster University
2019-2020
Harvard University
2009-2016
Harvard University Press
2014
Duke University
2006-2009
The University of Texas MD Anderson Cancer Center
2004-2005
Preoperative chemoradiotherapy may increase the R0 (curative) resection rate, overall survival (OS) duration, and disease-free (DFS) duration. We evaluated paclitaxel-based induction chemotherapy in patients with localized gastric or gastroesophageal adenocarcinoma to determine its feasibility, impact on type of pathologic response, OS, DFS.Patients operable, gastric, were eligible. Staging included endoscopic ultrasonography (EUS) laparoscopy. Patients received two 28-day cycles...
Risk assessment of rare natural hazards, such as large volcanic block and ash or pyroclastic flows, is addressed. Assessment approached through a combination computer modeling, statistical extreme-event probability computation. A model the hazard used to provide needed extrapolation unseen parts space. Statistical modeling available data determine initializing distribution for exercising model. In dealing with events, direct simulations involving are prohibitively expensive. The solution...
There is increasing appetite for analysing populations of network data due to the fast-growing body applications demanding such methods. While methods exist provide readily interpretable summaries heterogeneous populations, these are often descriptive or ad hoc, lacking any formal justification. In contrast, principled analysis results difficult relate back applied problem interest. Motivated by two complementary examples, we develop a Bayesian framework appropriately model complex while...
The value of baseline positron emission tomography (PET) for predicting overall survival (OS) or disease-free (DFS) is unclear in patients with nondistant metastatic (locoregional only) esophageal carcinoma. authors tested the hypothesis that, this setting, number PET abnormalities (NPA) would correlate OS and DFS.The current study analyzed localized carcinoma (Stages II III) who had a endoscopic ultrasonography (EUS) were all treated chemoradiotherapy followed by surgery. standardized...
This article introduces a new class of models for multiple networks. The core idea is to parameterize distribution on labeled graphs in terms Fréchet mean graph (which depends user-specified choice metric or distance) and parameter that controls the concentration this about its mean. Entropy natural such control, varying from point mass concentrated itself uniform over all given vertex set. We provide hierarchical Bayesian approach exploiting construction, along with straightforward...
The increasing prevalence of relational data describing interactions among a target population has motivated wide literature on statistical network analysis. In many applications, may involve more than two members the and this is appropriately represented by hypergraph. paper, we present model for hypergraph that extends well-established latent space approach graphs and, drawing connection to constructs from computational topology, develop whose likelihood inexpensive compute. A delayed...
We introduce a novel parameterization of distributions on hypergraphs based the geometry points in Rd. The idea is to induce by placing priors point configurations via spatial processes. This specification then used infer conditional independence models, or Markov structure, for multivariate distributions. approach results broader class models beyond standard graphical models. Factorizations that cannot be retrieved graph are possible. Inference nondecomposable possible without requiring...
Consider a population of individuals and network that encodes social connections among them. We are interested in making inference on finite super-population estimands function both individuals' responses the network, from sample. Neither sampling frame nor available. However, mechanism implicitly leverages to recruit individuals, thus partially revealing interactions sample, as well their responses. This is common setting arises, for instance, epidemiology healthcare, where samples...
The increasing availability of multiple network data has highlighted the need for statistical models heterogeneous populations networks. A convenient framework makes use metrics to measure similarity between In this context, we propose a novel Bayesian nonparametric model that identifies clusters networks characterized by similar connectivity patterns. Our approach relies on location-scale Dirichlet process mixture centered Erd\H{o}s--R\'enyi kernels, with components parametrized unique...
3507
3507 Background: Published analyses concluded that pathologic assessment of lymph nodes in colorectal cancer surgical specimens is often suboptimal and examining greater numbers increases the likelihood proper staging stage II tumors. The number may reflect technique thoroughness dissection. We postulated size addition to can impact survival. Methods: evaluated 129 consecutive patients with available materials adenocarcinoma colon rectum treated surgery alone at University Texas M. D....
We provide a general framework for constructing probability distributions on Riemannian manifolds, taking advantage of area-preserving maps and isometries. Control over distributions' properties, such as parameters, symmetry modality yield family flexible that are straightforward to sample from, suitable use within Monte Carlo algorithms latent variable models, autoencoders. As an illustration, we empirically validate our approach by utilizing proposed variational autoencoder space network...
The increasing prevalence of relational data describing interactions among a target population has motivated wide literature on statistical network analysis. In many applications, may involve more than two members the and this is appropriately represented by hypergraph. paper, we present model for hypergraph which extends well established latent space approach graphs and, drawing connection to constructs from computational topology, develop whose likelihood inexpensive compute. A...
A parametrization of hypergraphs based on the geometry points in $\mathbf{R}^d$ is developed. Informative prior distributions are induced through this by priors point configurations via spatial processes. This specification used to infer conditional independence models or Markov structure multivariate distributions. Specifically, we can recover both junction tree factorization as well hyper law. approach offers greater control distribution graph features than Erd\"os-R\'enyi random graphs,...