지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 자료유형
- 학위논문
- 저자정보
- 지도교수
- 이제영
- 발행연도
- 2020
- 저작권
- 영남대학교 논문은 저작권에 의해 보호받습니다.
이용수4
이 논문의 연구 히스토리 (2)
- 이 논문의 후속연구가 궁금하신가요?
- 연관 학술논문 또는 학술발표를 통해 보다 발전된 연구결과를 확인하실 수 있습니다.
- 이 논문의 연구 히스토리 확인하기
초록· 키워드
오류제보하기
In this study, we construct a nomogram to predict hypertension using complex samples.
Rao-Scott chi-squared test was used for the risk factors used to construct the nomogram, and
statistical models were modeled as logistic regression and naive Bayesian classifiers considering
the design effects of complex sample (stratication, clustering, sample weight, etc.). The
data used were the Korean national health and nutrition examination survey (KNHANES) 2013-2016. We identify risk factors for hypertension, and propose logistic nomogram and
Bayesian nomogram. In the case of logistic nomogram, age was the biggest risk factor,
followed by BMI, stroke, family history of hypertension and dyslipidemia. In the case of
Bayesian nomogram, age was the biggest risk factor for hypertension, followed by stroke,
diabetes, dyslipidemia and BMI. To compare the two nomograms, the left-aligned method
was applied to the Bayesian nomogram, and a similar prediction probability was verified in
the actual example. The AUC of ROC in logistic nomogram was 0.8301, 0.8286, and the R2
in calibration plot were 0.9146, 0.9002, respectively. The AUC of ROC in Bayesian nomogram
was 0.8169, 0.8168, and the R2 in calibration plot were 0.9340, 0.9235, respectively.
Therefore, the two nomograms were significantly reliable. Using the proposed nomogram,
it is expected that the treatment rate will increase as well as help in the recognition and
prediction of hypertension.
Rao-Scott chi-squared test was used for the risk factors used to construct the nomogram, and
statistical models were modeled as logistic regression and naive Bayesian classifiers considering
the design effects of complex sample (stratication, clustering, sample weight, etc.). The
data used were the Korean national health and nutrition examination survey (KNHANES) 2013-2016. We identify risk factors for hypertension, and propose logistic nomogram and
Bayesian nomogram. In the case of logistic nomogram, age was the biggest risk factor,
followed by BMI, stroke, family history of hypertension and dyslipidemia. In the case of
Bayesian nomogram, age was the biggest risk factor for hypertension, followed by stroke,
diabetes, dyslipidemia and BMI. To compare the two nomograms, the left-aligned method
was applied to the Bayesian nomogram, and a similar prediction probability was verified in
the actual example. The AUC of ROC in logistic nomogram was 0.8301, 0.8286, and the R2
in calibration plot were 0.9146, 0.9002, respectively. The AUC of ROC in Bayesian nomogram
was 0.8169, 0.8168, and the R2 in calibration plot were 0.9340, 0.9235, respectively.
Therefore, the two nomograms were significantly reliable. Using the proposed nomogram,
it is expected that the treatment rate will increase as well as help in the recognition and
prediction of hypertension.
목차
1. Introduction 12. Methodology 52.1. Rao-Scott chi-squared test 52.2. Logistic regression model with complex sample 62.3. Naive Bayesian classifier model with complex sample 82.4. Nomogram construction method 112.4.1. Nomogram construction of logistic regression model 122.4.2. Nomogram construction of naive Bayesian classifier model 132.4.3. Left-aligned method of nomogram for naive Bayesian classifier model 142.5. Nomogram validation method 163. Applications 163.1. complex sample materials 163.2. Rao-Scott chi-squared test results 193.3. Nomogram for hypertension using logistic regression model with complex sample 203.4. Nomogram for hypertension using naive Bayesian classifier model with complex sample 223.5. Comparison of logistic nomogram and left-aligned Bayesian nomogram 263.6. Validation for nomograms 284. Conclusions and discussions 32References 35
최근 본 자료
댓글(0)
0