Inteligentno sučelje za unos medicinskih podataka

scientific master's theses

• Title (english): Intelligent Interface for Medical Data Entry
• Author: Krešimir Šimunić
• Mentor: Damir Kalpić (mentor)
• Committee member: Vedran Mornar (član povjerenstva)
• Committee member: Damir Kalpić (član povjerenstva)
• Committee member: Vlatko Čerić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Applied Mathematics); Zagreb
• Defense date and country: 2000-06-07, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

Uobičajeni unos podataka temeljem izbora iz zatvorenog skupa, a uz uporabu standardnih izbornika, u medicini je neprikladan zbog velike kardinalnosti takvih skupova. Međutim, ako bi se povećala učinkovitost izbornika, njihova uporaba imala bi dodatnih prednosti u odnosu na unos slobodnog teksta. Apriorna vjerojatnost da će se pacijentu propisati neki od mnoštva lijekova nije velika, ali ako se zna od čega pacijent boluje, aposteriorna se vjerojatnost za određene lijekove znatno povećava. Podijeljeni izbornik je sučelje koje pokazuje nekoliko najvjerojatnijih stavki na vrhu liste (inteligentna lista), a potpuni skup stavki u preostalom dijelu liste. Iz literature znamo da su razvijena i analizirana 4 modela strojnog učenja (frekvencijski model, Bayesov model, model linearne regresije, model diskriminantne analize) za oblikovanje inteligentne liste. U radu je obrađena teoretska pozadina modela i analizirana karakteristika tih izbornika. Kao potencijalno bolji model, razvijen je model logističke regresije te inteligentni izbornik baziran na modelu logističke regresije i Bayesovom modelu. Uspoređuju se karakteristike tih modela. Dane su smjernice daljnjih istraživanja.

Abstract (english)

Due to the vast size of medical concepts, their entry from conventional menus is obviously problematic. If possible, however, menu-based entry has advantages over free-text processing. In the medical domain, there is a possibility to use conditional probability. A priori probability for prescription of a certain drug from the set of drugs is not high, but if we know patient’s problems, a posteriori probability can significantly increase. A split menu is an interface that shows a few likely options at the top of the menu, as an intelligent list and the complete set of options in the lower portion of the list. It is known from the literature that 4 machine learning models (the frequency model, Bayes model, linear regression model, and discriminant analysis model) have been developed for intelligent list creation. The thesis deals with the theoretical background of these models and it analyses the performances of these menus. As the method which may out-perform Bayes model, logistic regression model and a split menu based on logistic regression model and Bayes model have been developed. Performances of these models have been compared. A proposal for the follow-up in this particular research area has been made.

• Keywords: Medicinska informatika; podijeljeni izbornici; sučelje čovjek-računalo; inteligentna lista; izbornički unos podataka
• Keywords (english): Medical informatics; split menus; human-computer interface; intelligent hot list; menu data-entry
• Language: croatian
• URN:NBN: urn:nbn:hr:168:322521
• Study programme: Title: Postgraduate master programe in computing; Study programme type: university; Study level: postgraduate; Academic / professional title: magistar inženjer računarstva (magistar inženjer računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-04-04 08:04:05