A5032696713- Topic Modeling
- Natural Language Processing Techniques
- Advanced Text Analysis Techniques
- Semantic Web and Ontologies
- Sentiment Analysis and Opinion Mining
- Service-Oriented Architecture and Web Services
- Web Data Mining and Analysis
- Speech and dialogue systems
- Substance Abuse Treatment and Outcomes
- Biomedical Text Mining and Ontologies
- Neuroscience and Neuropharmacology Research
- Health Policy Implementation Science
- Amino Acid Enzymes and Metabolism
- Data Quality and Management
- Primary Care and Health Outcomes
- Geriatric Care and Nursing Homes
- Alcohol Consumption and Health Effects
- Chronic Disease Management Strategies
- Healthcare Policy and Management
- scientometrics and bibliometrics research
- Toxoplasma gondii Research Studies
- Climate Change and Health Impacts
- Machine Learning in Bioinformatics
- Pancreatic and Hepatic Oncology Research
- Frailty in Older Adults
Virginia Commonwealth University
2023-2024
University of Sheffield
2008-2020
University of Toledo
2019-2020
University of Cincinnati
2018
L3S Research Center
2014
St. Jude Children's Research Hospital
2009
University Medical Centre Mannheim
1994-2006
Heidelberg University
1994-2006
University Hospital Heidelberg
1994-2006
Twitter is the largest source of microblog text, responsible for gigabytes human discourse every day. Processing text difficult: genre noisy, documents have little context, and utterances are very short. As such, conventional NLP tools fail when faced with tweets other text. We present TwitIE, an open-source pipeline customised to at stage. Additionally, it includes Twitter-specific data import metadata handling. This paper introduces each stage TwitIE pipeline, which a modification GATE...
Abstract Converging evidence suggests bioenergetic defects contribute to the pathophysiology of schizophrenia and may underlie cognitive dysfunction. The transport metabolism lactate energetically couples astrocytes neurons supports brain bioenergetics. We examined concentration in postmortem (dorsolateral prefrontal cortex) subjects with schizophrenia, two animal models GluN1 knockdown mouse model mutant disrupted 1 (DISC1) model, as well inducible pluripotent stem cells (iPSCs) from a...
More than ever before, business analysts have access to public forums in which opinions and sentiments about companies, products, policies are expressed unstructured form. Mining information from sources is of great importance many intelligence applications such as credit rating or company reputation.
Background: End-of-life care (EOLC) discussions and decisions are common in pediatric oncology. Interracial differences have been identified adult EOLC preferences, but the relation of race to oncology has not reported. We assessed whether (white, black) was associated with frequency do-not-resuscitate (DNR) orders, number timing discussions, or among patients treated at our institution who died. Methods: reviewed records 380 died between July 1, 2001 February 28, 2005. χ2 Wilcoxon rank-sum...
The web and the social play an increasingly important role as information source for Members of Parliament their assistants, journalists, political analysts researchers. It provides crucial background information, like reactions to events comments made by general public. case study presented in this paper is driven two European parliaments (the Greek Austrian parliament) targets effective exploration archives. In paper, we describe semantic technologies deployed ease archived content present...
In this paper, we describe a set of reusable text processing components for extracting opinionated information from social media, rating it interestingness, and detecting opinion events. We have developed applications in GATE to extract named entities, terms events detect opinions about them, which are then used as the starting point event detection. The aggregated over larger sections text, give some overall sentiment topics documents, also degree interestingness based on diversity. go...
We present a supervised approach to automatically labelling topic clusters of reader comments online news.We use feature set that includes both features capturing properties local the cluster and capture aspects from news article outside cluster.We evaluate in an automatic manual, task-based setting.Both evaluations show outperform baseline method, which uses tf*idf select comment-internal terms for as labels.We illustrate how labels can be used generate summaries two alternative summary...
In this paper, we describe a set of reusable text processing components for extracting opinionated information from social media, rating it interestingness, and detecting opinion events. We have developed applications in GATE to extract named entities, terms events detect opinions about them, which are then used as the starting point event detection. The aggregated over larger sections text, give some overall sentiment topics documents, also degree interestingness based on diversity. go...
Abstract Background In silico data exploration is a key first step of exploring research question. There are many publicly available databases and tools that offer appealing features to help with such task. However, applications lack exposure or constrained unfriendly outdated user interfaces. Thus, it follows there resources relevant investigation medical disorders underutilized. Results We developed an R Shiny web application, called Kaleidoscope, address this challenge. The application...
While text mining and NLP research has been established for decades, there remain gaps in the literature that reports use of these techniques building real-world applications. For example, they typically look at single sometimes simplified tasks, do not discuss in-depth data heterogeneity inconsistency is common problems or their implication on development methods. Also, few prior work focused healthcare domain. In this work, we describe an industry project developed solutions to mine...
Abstract Background: Recruiting underrepresented people and communities in research is essential for generalizable findings. Ensuring representative participants can be particularly challenging practice-level dissemination implementation trials. Novel use of real-world data about practices the they serve could promote more equitable inclusive recruitment. Methods: We used a comprehensive primary care clinician practice database, Virginia All-Payers Claims Database, HealthLandscape mapping...
The Centers for Medicare & Medicaid Services (CMS) grades nursing home performance in antipsychotic prescribing quarterly, publishing findings as a quality measure. While scores have improved since 2011, marked variation between facilities persists. To assess gap changes best- and worst-performing deciles, we compared quarterly these groups pre-pandemic (April 2011 to March 2020) during the pandemic 2020 2022). Antipsychotic measure scores, improving pre-pandemic, deteriorated pandemic....