Yöntem ve uygulama açısından klinik karar destek sistemleri
No Thumbnail Available
Date
2013
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Open Access Color
OpenAIRE Downloads
OpenAIRE Views
Abstract
Bu tezde yapılan çalışmanın temel amacı Hastane Bilgi Yönetim Sistemlerine (HBYS) entegre olarak kullanılmakta olan Klinik Karar Destek Sistemlerinin (KKDS) yöntemsel açıdan incelenmesi ve temel sınıflandırma algoritmaları kullanarak KKDS performansının analiz edilmesidir. KKDS?ler; sağlık personeline alacağı kararlarda destek sağlayan, hekimlere, hastalara ait özel klinik bilgileri dikkate alarak karar verebilmeleri için yardım eden bilgisayar programlarıdır. Literatürdeki çalışmalar KKDS?lerin; sağlık bakım hizmetlerinin kalitesinin arttırılması, hastalıkların daha erken teşhis edilebilmesi, medikal hataların önlenmesi, hastalara uygun tedavi verilmesi ve maliyetlerin azaltılması konularında büyük faydalar sağladığını göstermektedir. Bu bağlamda, Türkiye?deki en büyük hastane zincirlerinden birisi olan Acıbadem Hastanesi Bilgi Yönetim Sistemi ve içerisinde yer alan KKDS?ler incelenmiştir. Bunun yanında örnek uygulama olarak UCI (University of California at Irvine) veritabanından elde edilen dermatoloji veri kümesi üzerinde k-En Yakın Komşu (K-nn), Naïve Bayes, Karar Ağacı ve Çok Katmanlı Algılayıcı (Multi Layer Perception - MLP) sınıflandırma algoritmalarının başarım oranları test edilmiştir. Sonuçlar; doğruluk, doğru pozitif (DP), yanlış pozitif (YP) ve alıcı işlem karakteristikleri (ROC) alanı kriterlerine göre değerlendirildiğinde Naïve Bayes algoritmasının daha iyi sonuç verdiği belirlenmiştir.
The main objective of this study is to examine Clinical Decision Support Systems (CDSS) integrated in Hospital Information Systems (HIS). The study also includes an implementation, for performance evaluation of classification algorithms applied in real world data set in the medical domain of dermatology. CDSS are computer programs that provide support for health professionals in taking decision using patients? clinical data. CDSS studies in the literature indicate substantial benefits such as improving the quality of health care services, the early diagnosis of diseases, medical errors prevention, appropriate treatment given to patients, and offers great benefits on reducing costs. In this context, Acıbadem HIS and CDSS used within the hospital was investigated. The implementation on dermatology data set obtained from UCI (University of California at Irvine) repository includes performance rates for k-Nearest Neighbor (K-nn), Naïve Bayes, Decision Tree and Multi Layer Perception (MLP) classification algorithms. The results were evaluated according to accuracy rate, true positive (TP), false positive (FP) and Receiver Operating Characteristics (ROC) area values. Consequently, Naïve Bayes algorithm results showed better performance results according to other competing classification algorithms.
The main objective of this study is to examine Clinical Decision Support Systems (CDSS) integrated in Hospital Information Systems (HIS). The study also includes an implementation, for performance evaluation of classification algorithms applied in real world data set in the medical domain of dermatology. CDSS are computer programs that provide support for health professionals in taking decision using patients? clinical data. CDSS studies in the literature indicate substantial benefits such as improving the quality of health care services, the early diagnosis of diseases, medical errors prevention, appropriate treatment given to patients, and offers great benefits on reducing costs. In this context, Acıbadem HIS and CDSS used within the hospital was investigated. The implementation on dermatology data set obtained from UCI (University of California at Irvine) repository includes performance rates for k-Nearest Neighbor (K-nn), Naïve Bayes, Decision Tree and Multi Layer Perception (MLP) classification algorithms. The results were evaluated according to accuracy rate, true positive (TP), false positive (FP) and Receiver Operating Characteristics (ROC) area values. Consequently, Naïve Bayes algorithm results showed better performance results according to other competing classification algorithms.
Description
Keywords
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Computer Engineering and Computer Science and Control
Turkish CoHE Thesis Center URL
Fields of Science
Citation
WoS Q
Scopus Q
Source
Volume
Issue
Start Page
End Page
94