J 2017

Finding overlapping terms in medical and health care curriculum using text mining methods: rehabilitation representation – a proof of concept

KAROLYI, Matěj, Martin KOMENDA, Radka JANOUŠOVÁ, Martin VÍTA, Daniel SCHWARZ et. al.

Basic information

Original name

Finding overlapping terms in medical and health care curriculum using text mining methods: rehabilitation representation – a proof of concept

Authors

KAROLYI, Matěj, Martin KOMENDA, Radka JANOUŠOVÁ, Martin VÍTA and Daniel SCHWARZ

Edition

MEFANET Journal, Brno, Facta Medica, 2017, 1805-9163

Other information

Language

English

Type of outcome

Článek v odborném periodiku

Field of Study

50300 5.3 Education

Country of publisher

Czech Republic

Confidentiality degree

není předmětem státního či obchodního tajemství

References:

URL

Organization unit

Faculty of Medicine

Keywords in English

natural language processing; curriculum; education; medical; rehabilitation

Tags

curriculum, Education, natural language processing, OPTIMED, rehabilitation

Tags

International impact, Reviewed
Změněno: 24/1/2017 09:54, Mgr. Matěj Karolyi

Abstract

V originále

Background: Various institutions dealing with higher medical and healthcare education have different methods of organising their study programmes, which typically involve hundreds of theoretically and clinically focused courses. The importance of a well-balanced curriculum is indisputable – the society needs qualified doctors because people’s health is necessary for the functioning and development of the entire society. Objectives: In this paper, we introduce our innovative approach to identify overlaps among medical or healthcare disciplines using term similarity. A close attention is focused on the discipline of Rehabilitation and Physical Medicine and its role in the General Medicine study field in the Faculty of Medicine at Masaryk University. Methods: Data and text mining techniques were used in practice, in accordance with a time-tested methodological background, which systematically covers all fundamental steps to discover and to extract knowledge from data repositories. In order to extract term similarities from a medical curriculum dataset, the CRISP-DM reference model was chosen as a well-documented practical guideline. Results: The achieved results clearly demonstrate overlapping areas among the defined disciplines in the explored curriculum. The resulting comprehensive analytical report presents the term occurrence in a set of figures and tables, which were thoroughly evaluated by experts familiar with the curriculum design process. Conclusions: In this case study, we have proposed an innovative method for identifying overlaps of terms occurring in medical and healthcare disciplines when compared to the discipline of Rehabilitation and Physical Medicine. The first results are promising in the sense of face validity. We believe that this approach can be used similarly to gain an objective overview of the entire curriculum.
Displayed: 17/11/2024 07:04