인두통과시간을 이용한 침습과 흡인의 예측 : 신경망 네트워크와 ROC curve를 중심으로

• 영문명: Prediction of Penetration and Aspiration by Pharyngeal Transit Time : Based on MLP & RBF Neural Network, ROC curve Analysis
• 발행기관: 대한치료과학회
• 저자명: 전제표 김현기 황승배 고명환
• 간행물 정보: 『대한치료과학회지』Vol. 6, No. 2, 56~60쪽, 전체 5쪽
• 주제분류: 의약학 > 재활의학
• 파일형태: PDF
• 발행일자: 2014.12.01

영문 초록

Objective: The purpose of this study is set to standard about prediction of penetration and aspiration by pharyngeal transit time when food through pharyngeal. Method: This study was to measure time when the swallowing type passes from the ramus of mandible to the upper esophagal sphincter, using the videofloroscopic imaging equipment as an object of 19 patients who had diseases with the basal ganglia. To investigate the effect of pharyngeal transit time on the penetration and aspiration, this study was to execute the bivariate logistic regression analysis. Furthermore, this study was to execute ROC curve analysis by considering the testing variable as pharyngeal transit time in order to set the standard value of inspection method for the purpose of diagnosis to distinguish between the penetration and aspiration. Results: This study has found out that the pharyngeal transit time had a significant effect on the penetration and aspiration statistically(p<.03). As a result of ROC curve analysis by considering the testing variable as pharyngeal transit time in order to set the standard value of inspection method for the purpose of diagnosis to predict the penetration and aspiration, the value of AUC was excellent as .90(p<.005). Conclusion: When it makes the multilayer neural networks and polynomial neural networks repeat training sufficiently, it comes to be similar to the value of classification accuracy of logistic regression analysis.

목차

I. 서론

II. 연구방법

III. 연구결과

IV. 고찰

V. 결론

키워드

Penetration Aspiration Pharyngeal Transit Time