http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Risk Prediction of Hypertension Complicates using Classification Techniques
Wonji Lee(이원지),Junghye Lee(이정혜),Hyeseon Lee(이혜선),Chi-Hyuck Jun(전치혁),Il-su Park 대한산업공학회 2014 대한산업공학회 춘계학술대회논문집 Vol.2014 No.5
A hypertension complications is one of the sources causing the national medical expenditures to increase. We aim to score the risk of hypertension complications for hypertension patients, using national healthcare database established by Korean National Health Insurance Corporation. We apply classification techniques such as logistic regression, linear discriminant analysis and classification and regression tree to score the risk of hypertension complication onset and also compare the performance of those methods. These three methods seem to perform similarly although the logistic regression performs better than others marginally. This study is meaningful in that the database used is a representative sample for the whole nation.
A modified AdaBoost with SVM for relaxing class imbalance problem
Wonji Lee(이원지),Chi-Hyuck Jun(전치혁) 대한산업공학회 2015 대한산업공학회 춘계학술대회논문집 Vol.2015 No.4
Class imbalance problem is a challenging issue in data mining and machine learning. With imbalanced data set, the object of interest is usually a minority class, but standard classifiers tend to be poor at classifying a minority class. Ensemble method such as AdaBoost is a representative approach known for having best performance to relax imbalance problem. The objective is to propose a modified AdaBoost model with SVM component focusing class imbalance problem by categorizing instances and differentiating instance weights according to the category of instances.
A modified AdaBoost with SVM for relaxing class imbalance problem
Wonji Lee(이원지),Chi-Hyuck Jun(전치혁) 한국경영과학회 2015 한국경영과학회 학술대회논문집 Vol.2015 No.4
Class imbalance problem is a challenging issue in data mining and machine learning. With imbalanced data set, the object of interest is usually a minority class, but standard classifiers tend to be poor at classifying a minority class. Ensemble method such as AdaBoost is a representative approach known for having best performance to relax imbalance problem. The objective is to propose a modified AdaBoost model with SVM component focusing class imbalance problem by categorizing instances and differentiating instance weights according to the category of instances.
Risk Prediction of Hypertension Complicates using Classification Techniques
Wonji Lee(이원지),Junghye Lee(이정혜),Hyeseon Lee(이혜선),Chi-Hyuck Jun(전치혁),Il-su Park 한국경영과학회 2014 한국경영과학회 학술대회논문집 Vol.2014 No.5
A hypertension complications is one of the sources causing the national medical expenditures to increase. We aim to score the risk of hypertension complications for hypertension patients, using national healthcare database established by Korean National Health Insurance Corporation. We apply classification techniques such as logistic regression, linear discriminant analysis and classification and regression tree to score the risk of hypertension complication onset and also compare the performance of those methods. These three methods seem to perform similarly although the logistic regression performs better than others marginally. This study is meaningful in that the database used is a representative sample for the whole nation.
묘사와 서술을 통한 건축 읽기 - 부전교회와 양덕성당을 중심으로 -
이원지(WonJi Lee),백승한(Seunghan Paek) 대한건축학회 2022 대한건축학회 학술발표대회 논문집 Vol.42 No.1
Not only matters consist architectures. Essence of architectures are beyond them. To prosper the true architectures with users of architectures, in practical ways, not merely in academic ways, communicating the essence should be analysed in detail and logic. Users of architectures could accept two things. The first thing is optical depiction which is a sort of sense. The last is spatial description that contains arrangement of space and view of world by architects. This study epitomizes Bujeon Presbyterian Church in Busan and Yangduk Cathedral in Jinhae, which is because religious architectures are relatively clear in purpose of use. By studying them, the study would show how to proceed public-friendly design process in architecture, which eventually make architectural culture mature and prosper.
Hypertension Occurrence Analysis using a Bayesian Network
Junghye Lee(이정혜),Wonji Lee(이원지),Hyeseon Lee(이혜선),Chi-Hyuck Jun(전치혁) 대한산업공학회 2014 대한산업공학회 춘계학술대회논문집 Vol.2014 No.5
The Bayesian network (BN) is a useful method for modeling healthcare issues since a BN can graphically represent causal relationships among variables and provide its probabilistic information,. In this study, we apply a BN method to hypertension occurrence analysis. This study used the National Health Insurance Corporation (NHIC) database from 2002 to 2010 which contains more than 100,000 cases of personal medical examinations in Korea. We investigate the causality for hypertension occurrence by a structure learning step, and then evaluate the performance to predict hypertension occurrence through parameter learning and inference steps. It is shown that the BN outperforms other prediction methods such as logistic regression, naive Bayes and support vector machine in terms of sensitivity. In addition, the BN has advantages in interpreting which variables affect the hypertension occurrence and how they are related to each other.
Hypertension Occurrence Analysis using a Bayesian Network
Junghye Lee(이정혜),Wonji Lee(이원지),Hyeseon Lee(이혜선),Chi-Hyuck Jun(전치혁) 한국경영과학회 2014 한국경영과학회 학술대회논문집 Vol.2014 No.5
The Bayesian network (BN) is a useful method for modeling healthcare issues since a BN can graphically represent causal relationships among variables and provide its probabilistic information,. In this study, we apply a BN method to hypertension occurrence analysis. This study used the National Health Insurance Corporation (NHIC) database from 2002 to 2010 which contains more than 100,000 cases of personal medical examinations in Korea. We investigate the causality for hypertension occurrence by a structure learning step, and then evaluate the performance to predict hypertension occurrence through parameter learning and inference steps. It is shown that the BN outperforms other prediction methods such as logistic regression, naive Bayes and support vector machine in terms of sensitivity. In addition, the BN has advantages in interpreting which variables affect the hypertension occurrence and how they are related to each other.