HOMI17-StrategiesForKnowledgeAcquisition

Descrição

informaticsBoards NotForReview Notas sobre HOMI17-StrategiesForKnowledgeAcquisition, criado por Michael Riben em 20-08-2013.
Michael Riben
Notas por Michael Riben, atualizado more than 1 year ago
Michael Riben
Criado por Michael Riben mais de 10 anos atrás
76
0

Resumo de Recurso

Página 1

Medical Knowledge-Base = A systematically organized collection of medical knowledge that is accessible electronically and interpretable by the computer. Usually contains a lexicon or vocabulary of allowed terms and specific relationships between terms in the lexicon.Meta-analysis - a discipline that critically reviews and statistically combines results of previous research Knowledge Base construction on literatures requires a quality assessment ofthe sources of information used to create the knowledgebase. 2 categories for knowledge acquisitions:1. Expert utilized knowledge acquisitions editors to build knowledge2. Literature is used and reviewed by folks to identify critical knowledge for the knowledge baseKnowledge editors are critical for developing a tool to not require a computer scientist to work so close with the medical expertsLiterature based acquisition is slow- 10 -20 days full time to build 1 disease profile in QMRKnowledge Acquisition Trends:1) creation of knowledge expert systems by domain 2) High level representation3) better user interfaces4) Natural Language Making KB's Reproducible Explicitly stated purpose Carefully defined components of the KB Reliable sources of information used Methods for resolving conflicts must be applied when separate sources information differs Clearly delineated procedures for testing, updating, --> tested over time Teiresias - Developed by Davis- developed for the rule expert system Mycin = first to do computer assisted  knowledge acquisition, separated knowledge and meta-knowledge. Meta knowledge was contained in "rule models"which were compact descriptions of structure, organization, and content of a group of related rules.  - wouldn't work with Backward chaining expert systemsOpal - graphical knowlege acquisition sstem for ONCOCIN cancer therapy management program - precursor to Protege , point and click forms, flow chart procedure mapping , strong assumptions about the domain. generates a fixed problem solving modle and tailored for the oncocin framework. Protege 1  - abstracted a layer above opal so that could be used to generate knowledge acquisition tool , domain independent, originally Internest-1/QMR

New Page

Semelhante

HOMI15-MethodsforDecisionSupport
Michael Riben
HOMI19 Modeling Health CAre for IS Devolopment
Michael Riben
Information Processing
Michael Riben
Chapter6 -HOMI- coding/Classification
Michael Riben
HOMI-10 Image Processing and Analysis
Michael Riben
Chapter 14 Nursing InfoSystems
Michael Riben
Chapter 4 HOMI-Database Management
Michael Riben
Chapter 2 - HOMI
Michael Riben
Chapter8-HOMI-Biosignal Analysis
Michael Riben
Chapter 3 HOMI- Data Processing
Michael Riben
Chapter 5-HOMI-Telcom,Networks,Integration
Michael Riben