AI Article Synopsis
- A maximum entropy-based system was developed for identifying named entities in biomedical abstracts.
- The system achieved an F-score of 83.2% in the BioCreative evaluation and 70.1% in the BioNLP evaluation.
- Key features of the system include the use of local features, attention to boundary identification, and the integration of external resources, along with a discussion of data annotation issues that affected performance.
Article Abstract
We present a maximum entropy-based system for identifying named entities (NEs) in biomedical abstracts and present its performance in the only two biomedical named entity recognition (NER) comparative evaluations that have been held to date, namely BioCreative and Coling BioNLP. Our system obtained an exact match F-score of 83.2% in the BioCreative evaluation and 70.1% in the BioNLP evaluation. We discuss our system in detail, including its rich use of local features, attention to correct boundary identification, innovative use of external knowledge resources, including parsing and web searches, and rapid adaptation to new NE sets. We also discuss in depth problems with data annotation in the evaluations which caused the final performance to be lower than optimal.
Download full-text PDF |
Source |
|---|---|
| http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2448599 | PMC |
| http://dx.doi.org/10.1002/cfg.457 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Mesenchymal stromal cells promote the formation of lung cancer organoids via Kindlin-2.
Stem Cell Res Ther
January 2025
Shenzhen Key Laboratory of Epigenetics and Precision Medicine for Cancers, Department of Thoracic Surgery, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital & Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Shenzhen, 518116, China.
Background: Patient-derived lung cancer organoids (PD-LCOs) demonstrate exceptional potential in preclinical testing and serve as a promising model for the multimodal management of lung cancer. However, certain lung cancer cells derived from patients exhibit limited capacity to generate organoids due to inter-tumor or intra-tumor variability. To overcome this limitation, we have created an in vitro system that employs mesenchymal stromal cells (MSCs) or fibroblasts to serve as a supportive scaffold for lung cancer cells that do not form organoids.
View Article and Find Full Text PDFEvaluation of a national framework for rational use of medicines in Kazakhstan and its role in improving medicine use practices at the organizational and national levels.
BMC Health Serv Res
January 2025
Department of Clinical Pharmacology and Evidence-Based Medicine, NCJSC "Karaganda Medical University", 40, Gogolya St, Karaganda, 100000, Kazakhstan.
Background: Kazakhstan inherited the Semashko health system model, known for the centralized adoption of rules at the Ministry of Health (MoH) level that regulate the healthcare system. In 2019 MoH established a national framework with indicators aimed at collecting qualitative and quantitative data from healthcare organizations as part of their annual self-evaluation, and biannual external evaluation by the National Research Center for Health Development (NRCHD). The purpose of this study was to pilot the MoH framework on rational use of medicines and evaluate its effects on medicine use practices in health care organizations and at the national level.
View Article and Find Full Text PDFRacial, socioeconomic, and payer status disparities in utilization of unicompartmental knee arthroplasty in the USA.
Knee Surg Relat Res
January 2025
Department of Orthopedic Surgery, The Johns Hopkins University School of Medicine, 601 N Caroline St., Baltimore, MD, 21287, USA.
Background: Unicompartmental knee arthroplasty (UKA) is a surgical treatment for knee osteoarthritis associated with lower morbidity compared with total knee arthroplasty (TKA) in patients with isolated unicompartmental knee arthritis. As disparities have been noted broadly in arthroplasty care, it follows that such disparities might be present in the utilization of UKA relative to TKA. This study therefore examined racial/ethnic, socioeconomic, and payer status differences in utilization of UKA.
View Article and Find Full Text PDFUnderstanding barriers and facilitators to implementation of a patient safety bundle for pregnancy-related severe hypertension in 3 North Carolina outpatient clinics: a qualitative study.
Implement Sci Commun
January 2025
Department of Obstetrics and Gynecology, Division of Maternal and Fetal Medicine, University of North Carolina School of Medicine, Chapel Hill, NC, 27599, USA.
Background: Pregnancy related hypertension is a leading cause of preventable maternal morbidity and mortality in the US, with consistently higher rates affecting racial minorities. Many complications are preventable with timely treatment, in alignment with the Alliance for Innovation on Maternal Health's Patient Safety Bundle ("Bundle"). The Bundle has been implemented successfully in inpatient settings, but 30% of preeclampsia-related morbidity occurs in outpatient settings in North Carolina.
View Article and Find Full Text PDFDevelopment and pilot testing of INTERVENER, a web-based tool to match barriers to the cancer continuum organization to evidence-based interventions.
BMC Health Serv Res
January 2025
Early Detection, Prevention & Infections Branch, International Agency for Research on Cancer, 25 Avenue Tony Garnier, Lyon, 69366 Cedex 07, France.
Background: Barriers to the cancer continuum organization and interventions to approach them have been identified; however, there is a lack of a tool matching them. Our aim was to develop a web-based tool to identify the main barriers to the process of the cancer continuum organization, and propose matched evidence-based interventions (EBI) to overcome them.
Methods: A questionnaire on barriers at six steps of the process of the cancer continuum organization was answered by collaborators.
Want AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!