J48 Tree와 AD Tree를 이용한 원내 감염의 분석

• 영문명: Analysis of Nosocomial Infection using J48 Tree and AD Tree
• 발행기관: 한국IT마케팅학회
• 저자명: 박재현(Jae Hyun Park) 정용규(Yong Gyu Jung)
• 간행물 정보: 『한국IT마케팅학회 논문집』한국IT마케팅학회 논문집 1권 1호, 41~43쪽, 전체 3쪽
• 주제분류: 경제경영 > 경영학
• 파일형태: PDF
• 발행일자: 2014.11.27

영문 초록

(Introduction) In the 21st century world-aging, influenza, due to natural disasters, has increased people are looking for a hospital, thereby, hospital infection is rapidly increasing. If you look at the information that was published in 2005, and 5-10% of patients were admitted to the General Hospital nosocomial infection has occurred. These hospital-acquired infections, because it results in dropping the hospitals of credibility and reputation, many hospitals perform preliminary survey, attempts to prevent nosocomial infections. (literature study) In this paper, we want to analyze and take advantage of Decision Tree (decision tree) of data mining techniques that have this article. (experimental) Data to be used in the This paper, in the United States of University Hospital, representation and numerical patients who visit the hospital was expected medical team the probability of infection in CDI Community Onset and CLABSI Coronaty ICU, a number that was actually the onset as the data, by using this data, in this paper several tree decision tree, and analyzed using the most widely used J48 Tree and AD Tree. Before analyzing the data, the pre-treatment process, has been added a difference in the prediction data and the actual onset of the onset of data in the new atrribute. (Evaluation and discussion) As a result of the result which can be confirmed were analyzed by J48 Tree throughout the experiment results, accuracy and are seen to be slightly predominant than the accuracy of the results analyzed in AD Tree. Using these analytical techniques the results of analyzing the data, CLABSI Coronary ICU than that it has I~echuku less almost all are developing, CDI Community Onset was found to be developing less extent approximately half. (infected Number of actual observation) Data used in Conclusion In this paper, it can be known that it is better to be analyzed in J48 Tree analysis method, the results of the analysis of these data infections observed and Infections predicted When we compare the difference of (expected number of infected) with predicted probabilities CDI Community Onset is about 57%, CLABSI Coronary ICU was confirmed by about 94%. When I analogy Based on these results, the case of the University Hospital, to investigate these nosocomial bacteria, are expected to be able to Reduce the nosocomial infection via post-processing corresponding thereto.

목차

Ⅰ. 서론

Ⅱ. 문헌연구

Ⅲ. 실험

Ⅳ. 평가 및 토의

Ⅴ. 결론

참고문헌

키워드

군집화 병원 내 감염 예측 정확도 Decision Tree J48 Tree AD Tree