• 제목/요약/키워드: Conditional inference tree

검색결과 4건 처리시간 0.017초

제주 실시간 일사량의 기계학습 예측 기법 연구 (A Study on Prediction Techniques through Machine Learning of Real-time Solar Radiation in Jeju)

  • 이영미;배주현;박정근
    • 한국환경과학회지
    • /
    • 제26권4호
    • /
    • pp.521-527
    • /
    • 2017
  • Solar radiation forecasts are important for predicting the amount of ice on road and the potential solar energy. In an attempt to improve solar radiation predictability in Jeju, we conducted machine learning with various data mining techniques such as tree models, conditional inference tree, random forest, support vector machines and logistic regression. To validate machine learning models, the results from the simulation was compared with the solar radiation data observed over Jeju observation site. According to the model assesment, it can be seen that the solar radiation prediction using random forest is the most effective method. The error rate proposed by random forest data mining is 17%.

데이터마이닝 기법들을 통한 제주 안개 예측 방안 연구 (A Study on Fog Forecasting Method through Data Mining Techniques in Jeju)

  • 이영미;배주현;박다빈
    • 한국환경과학회지
    • /
    • 제25권4호
    • /
    • pp.603-613
    • /
    • 2016
  • Fog may have a significant impact on road conditions. In an attempt to improve fog predictability in Jeju, we conducted machine learning with various data mining techniques such as tree models, conditional inference tree, random forest, multinomial logistic regression, neural network and support vector machine. To validate machine learning models, the results from the simulation was compared with the fog data observed over Jeju(184 ASOS site) and Gosan(185 ASOS site). Predictive rates proposed by six data mining methods are all above 92% at two regions. Additionally, we validated the performance of machine learning models with WRF (weather research and forecasting) model meteorological outputs. We found that it is still not good enough for operational fog forecast. According to the model assesment by metrics from confusion matrix, it can be seen that the fog prediction using neural network is the most effective method.

Early diagnosis of jaw osteomyelitis by easy digitalized panoramic analysis

  • Park, Moo Soung;Eo, Mi Young;Myoung, Hoon;Kim, Soung Min;Lee, Jong Ho
    • Maxillofacial Plastic and Reconstructive Surgery
    • /
    • 제41권
    • /
    • pp.6.1-6.10
    • /
    • 2019
  • Background: Osteomyelitis is an intraosseous inflammatory disease characterized by progressive inflammatory osteoclasia and ossification. The use of quantitative analysis to assist interpretation of osteomyelitis is increasingly being considered. The objective of this study was to perform early diagnosis of osteomyelitis on digital panoramic radiographs using basic functions provided by picture archiving and communication system (PACS), a program used to show radiographic images. Methods: This study targeted a total of 95 patients whose symptoms were confirmed as osteomyelitis under clinical, radiologic, pathological diagnosis over 11 years from 2008 to 2017. Five categorized patients were osteoradionecrosis, bisphosphonate-related osteonecrosis of jaw (BRONJ, suppurative and sclerosing type), and bacterial osteomyelitis (suppurative and sclerosing type), and the control group was 117 randomly sampled. The photographic density in a certain area of the digital panoramic radiograph was determined and compared using the "measure area rectangle," one of the basic PACS functions in INFINITT PACS® (INFINITT Healthcare, Seoul, South Korea). A conditional inference tree, one type of decision making tree, was generated with the program R for statistical analysis with SPSS®. Results: In the conditional inference tree generated from the obtained data, cases where the difference in average value exceeded 54.49 and the difference in minimum value was less than 54.49 and greater than 12.81 and the difference in minimum value exceeded 39 were considered suspicious of osteomyelitis. From these results, the disease could be correctly classified with a probability of 88.1%. There was no difference in photographic density value of BRONJ and bacterial osteomyelitis; therefore, it was not possible to classify BRONJ and bacterial osteomyelitis by quantitative analysis of panoramic radiographs based on existing research. Conclusions: This study demonstrates that it is feasible to measure photographic density using a basic function in PACS and apply the data to assist in the diagnosis of osteomyelitis.

전술제대 공격작전간 전투원 생존성에 관한 연구 (Analysis of Survivability for Combatants during Offensive Operations at the Tactical Level)

  • 김재오;조형준;김각규
    • 응용통계연구
    • /
    • 제28권5호
    • /
    • pp.921-932
    • /
    • 2015
  • 본 연구에서는 증강된 보병대대의 과학화 전투훈련 데이터 중 공격작전에 관한 장병들의 생존분석을 실시하였다. 과학화 전투훈련은 KCTC(Korea Combat Training Center)로 불리는 전투훈련장에서 MILES(Multiple Integrated Lazer Engagement System)와 중앙통제장비체계 등 과학화된 훈련장비와 체계 운용하 훈련부대가 적 전술 및 무기체계를 사용하는 전문 대항군과 실시하는 쌍방 자유기동훈련이다. 이는 훈련기간 동안 훈련지역의 모든 데이터가 저장되어 훈련통제 뿐 아니라 분석 및 사후검토를 할 수 있는 첨단화된 군사 훈련으로 통계적 분석이 가능한 데이터를 제공한다. 분석방법은 모수적 분포 가정이 필요하지 않은 Cox의 비례위험모형을 적용하였으며, 보다 풍부하고 용이한 해석을 위해 의사결정나무모형(CART(Classification and Regression Trees), GUIDE(Generalized, Unbiased, Interaction Detection and Estimation), CTREE(Conditional Inference Trees))을 활용하였다. Cox 비례위험모형의 비례성 가정을 확인하여 이를 위배하는 변수에 대해서 층화하여 분석하고, Cox 비례위험모형 결과 복무기간에 관한 해석이 용이하지 않아 단변량으로 local 회귀분석을 통해 추가적인 해석을 시도하였다. CART, GUIDE, CTREE는 모형의 특성별로 나무모형을 형성하며 이를 통하여 다양한 해석이 가능하다.