2009.10.30 Senator Jeong, Soo Kyung Lee, Hong-Gee Kim Biomedical Knowledge Engineering Lab.,

2009.10.30

Senator Jeong, Soo Kyung Lee, Hong-Gee KimBiomedical Knowledge Engineering Lab., Seoul National University1

PurposeAnalyze the knowledge structure of Korean medical informaticsin quantitative way.

QuestionsWhat are the important topics of Korean Medical Informatics?

What are the newly emerging research topics?

Method

Co-word AnalysisarticleTopic ATopic BThese two topics are likely to be relatedarticlearticleTopic CCo-word Analysis Workflow & Tool

BiKE Text Analyzer

Data Collection and Treatment

Data CollectionSourceThe Journal of Korean Society of Medical InformaticsKOSMI SymposiumsTime Coverage: 1995-2008 (14 years)Data Corpus: 1,075 papers915 papers (excluded abstract-free papers)

Data Treatment: Translatecorrected the English termsData Treatment: Variables

Term Extraction and NormalizationWords plural form singular formsynonyms controlledExtracted 2-5 gram terms as variablestotal number of n-gram terms=2,954The most frequently occurring terminformation system (term frequency=533). Occurrence Threshold: less than 5 timesTerm variables for analysis: 748

Term Variable Selection

Extract Terms, Term Frequency

Term Weight, Co-occurance

Term WeightTerm Similarity, Matrix File

Cosine Coefficient (0-1)Network Analyze and Visualization

Pajek

ResultsTop 100 research Topics(Tf5; N=100; Cosine>0.15; k-component.1; component=7)

Top 50 research Topics(Tf5; N=100; Cosine>0.15)

Important TopicsTop 50 research Topics of Korean Medical Informatics(Tf5; N=50; N=50; Cosine>0.15; k-component.1; component=9)

Top 100 global research topics in MI(tf10)

CDS*Info.SystemCPOEPACSCBIRPDACPGEMRMach.LearningEHRPHRSVMNLPSenator Jeong, Hong-Gee Kim. Intellectual Structure of Biomedical Informatics reflected in Scholarly Events. Scientometrics. 2009. [in Press]22

Interesting PhenomenaTop 50 research Topics of Korean Medical Informatics(Tf5; N=50; N=50; Cosine>0.15; k-component.1; component=9)

Research Topic TrendsNewly rising topicsIdentified the topics which represented the lowest 10% in the low frequency group in the preceding period(s), and which also remained in the highest 10% (5% in the years 207-2008) in the high frequency group in the following periods.

101998~20002001~2003, 2004~20062007-2008

ConclusionThe findings can be of help to decide which technologies and themes should be included in medical informatics curriculum to meet learners needs.

Q & AThank you