구강관리 번역(한국어 원본)이 연구는 일부 대학생 235명을 대상으로 구조화된 설문을 통하여 치통과 관련된 구강건강관리실태를 조사 분석하고, CART 모형을 이용하여 의사결정나무분석을 하였고, 신경망 모형을 이용하여 회귀분석과 비교한 치통의 발생 예측률을 제시하였다. 회귀분석 결과 현재 건강상태, 지난 1년간 건강상태, 규칙적인 식사여부, 칫솔질 사용시 잇몸출혈에서는 유의하지 않았다.
연구결과 대상자의 치통 발생 유무에 따른 현재 건강상태, 지난 1년간의 건강상태
는 통계적으로 차이가 있었다(p<0.05).
대상자의 식생활습관에 따른 치통발생 유무에서 규칙적인 식사여부는 통계적으로 차이가 있었다(p<0.05).
대상자의 구강건강관리습관에 따른 치통발생 유무에서는 칫솔질이나 치실사용시 잇몸 출혈여부에서 통계적으로 차이가 있었다(p<0.05).
CART 결과는 식사시간이 불규칙하면서 현재 건강상태가 나쁜 경우 치통 발생이 보통인 경우보다 높게 나타났다.
신경망 모형을 이용한 치통발생 예측 정확도는 72.8%이었다.
이상의 결과는 치통의 발생을 예측하는데 회귀분석에 비하여 신경망 모형이 매우 유용한 기법임을 시사하며, 향후 치통을 포함한 다른 구강 질환을 예측하는데 매우 도움이 될 것으로 기대된다. |
구강관리 번역(영어 번역본)This study has researched and analyzed the actual conditions of the toothache-related oral hygiene habits through a structured survey on 235 college students, followed by the decision tree analysis using the CART model, and has presented the predictive value for occurrence of toothache in comparison to the regression analysis using the neural network model. According to the result of the regression analysis, the predictive value was not significant in the present health condition, health condition for the past year, regular eating routine and occurrence or non-occurrence of bleeding during tooth-brushing.
The result showed a statistical difference (p<0.05) between the present health condition and the health condition of the past year depending on the occurrence or non-occurrence of toothache in the research subjects.
There was a statistical difference (p<0.05) between the presence or non-presence of regular eating routine in the occurrence or non-occurrence of toothache depending on the eating habits of the research subjects.
In terms of the occurrence or non-occurrence of toothache depending on the oral hygiene habits of the research subjects, there was a statistical difference (p<0.05) in the occurrence or non-occurrence of bleeding from gums during tooth-brushing or flossing.
According to the CART result, the occurrence of toothache was higher for those who answered to have irregular meal time and bad health condition than those who answered average to both of the surveyed questions.
The predictive accuracy of toothache occurrence using the neural network model was 72.8%.
The above results show that the neural network model is more useful technique than the regression analysis in predicting the occurrence of toothache, and thus the neural network model is expected to be very helpful in predicting not only toothache but also other oral diseases in the future research studies. |
