• 한국수자원학회논문집
• /
• 제51권7호
• /
• pp.575-584
• /
• 2018

최근 이상기후로 인한 집중호우 발생빈도와 이로 인한 국지적인 홍수 피해가 증가하고 있다. 이러한 점에서 홍수피해 예방측면에서 수치예보 정보 활용이 요구되고 있다. 그러나 수치예보모델은 초기 조건 및 지형적 요인으로 인해 시공간적 편의가 존재하며 실시간 예측정보로 활용하기 전에 모형결과에 대한 편의보정이 요구된다. 본 연구에서는 관측지점 기준으로 편의 보정계수를 산정하는 과정에서 모든 관측소간의 상관성을 거리의 함수로 고려하여 미계측지점의 편의 보정계수를 공간적으로 확장할 수 있는 Bayesian Kriging 기반 MFBC 기법을 개발하였다. 본 연구에서 개발한 방법은 미계측 유역에 대해서도 보정계수를 효과적으로 추정하는 것이 확인되었으며, 비교적 고해상도로 72시간(3일) 정도까지 예측강우 정보를 활용하는 것이 가능할 것으로 판단된다.

• Asian Pacific Journal of Cancer Prevention
• /
• 제17권6호
• /
• pp.2895-2899
• /
• 2016

Background: Increasing breast cancer (BC) incidence rates have been reported from India; causal factors for this increased incidence are not understood and diagnosis is mostly in advanced stages. Trivandrum exhibits the highest BC incidence rates in India. This study aimed to estimate trends in incidence by age from 2005-2014, to predict rates through 2020 and to assess the stage at diagnosis of BC in Trivandrum. Materials and Methods: BC cases were obtained from the Population Based Cancer Registry, Trivandrum. Distribution of stage at diagnosis and incidence rates of BC [Age-specific (ASpR), crude (CR) and age-standardized (ASR)] are described and employed with a joinpoint regression model to estimate average annual percent changes (AAPC) and a Bayesian model to estimate predictive rates. Results: BC accounts for 31% (2681/8737) of all female cancers in Trivandrum. Thirty-five percent (944/2681) are <50 years of age and only 9% present with stage I disease. Average age increased from 53 to 56.4 years (p=0.0001), CR (per $10^5$ women) increased from 39 (ASR: 35.2) to 55.4 (ASR: 43.4), AAPC for CR was 5.0 (p=0.001) and ASR was 3.1 (p=0.001). Rates increased from 50 years. Predicted ASpR is 174 in 50-59 years, 231 in > 60 years and overall CR is 80 (ASR: 57) for 2019-20. Conclusions: BC, mostly diagnosed in advanced stages, is rising rapidly in South India with large increases likely in the future; particularly among post-menopausal women. This increase might be due to aging and/or changes in lifestyle factors. Reasons for the increased incidence and late stage diagnosis need to be studied.

• 한국정보처리학회논문지
• /
• 제1권1호
• /
• pp.14-22
• /
• 1994

본 논문은 스미스의 베이지안 소프트웨어 신뢰도 성장모형을 기반으로 테스팅 단계에서의 소프트웨어 신뢰도에 대한 두가지 베이즈 추정량에 그에 대한 평가 알고 리즘을 제안하는데 목적이 있다. 그 방법으로 사전정보 클래스로서 일양사전분포보다 더 일반적인 베타사전분포 BE(a.b)를 사용하였다. 그 연구 과정으로 베이지안 추정절 차에 있어서 제곱오차결손함수와 해리스결손함수를 고려하고, 컴퓨터 시뮬레이션을 통 해서 소프트웨어 신뢰도에 대한 베이즈추정량들과 그에 따른 알고리즘을 이용하여 평 균자승오차 성능을 비교한다. 연구 결과로써 a가 크면 클수록 그리고 b가 적으면 적을 수록 해리스결손함수하의 소프트웨어 신뢰도의 베이즈추정량이 평균자승오차 성능의 관점에서는 더욱 유효하고, a 가 b보다 더 클 때 공액사전분포인 베타사전분포상의 소 프트웨어 신뢰도의 베이즈추정량이 비정보사전분포인 일양사전분포상에서 소프트웨어 신뢰도의 베이즈추정량보다는 성능이 더 좋다는 결론을 얻는다.

• 한국식품저장유통학회지
• /
• 제30권3호
• /
• pp.459-471
• /
• 2023

뱀장어 생산단계 안전성조사 부적합여부에 영향을 미치는 특성변수를 베이지안 네트워크(BN) 모델을 적용하여 분석하였다. 2012년부터 2021년까지의 통합식품안전정보망(IFSIN)의 뱀장어 생산단계 안전성조사 데이터에 양식장의 HACCP 정보, 지리적 정보 및 용수환경 데이터를 연계하여 BN 모델을 수립하였다. 뱀장어의 부적합여부에 영향을 주는 특성변수로 양식장의 HACCP 인증여부, 양식장의 이전 5년간 검사대상 유해물질의 검출여부, 해당 양식장의 이전 5년간 부적합적발이력, 사용되는 용수환경의 적정성이 제안되었으며, 이때 용수환경의 적정성은 총대장균군과 총유기탄소량으로부터 산출되었다. 뱀장어 부적합이 발생할 확률이 가장 높은 경우는 지난 5년간 검사대상 유해물질의 검출이력이 있으면서 동시에 부적합 적발 이력이 있는 HACCP 인증을 받지 않은 양식장으로서, 용수환경도 총대장균군 또는 총유기탄소가 높아 오염이 의심되는 용수를 사용하는 경우로 이때 부적합이 발생할 확률은 24.5%로 뱀장어 생산단계 안전성 조사 시 부적합률인 0.26%의 94배 높았다. 2022년 1월부터 8월까지 뱀장어 양식장 안전성조사 결과를 시험용 데이터세트(6,785건 중 부적합 15건)로 하여 BN 모델의 적정성을 검토하였다. 영향강도가 높았던 설명변수인 HACCP, 검출이력, 부적합이력으로 구성한 BN 모델을 시험용 데이터세트에 적용한 결과 부적합일 확률이 15.8%로 시험용데이터의 부적합률인 0.22%의 약 71.4배 개선할 수 있었다. 그러나 이 모델의 재현율은 0.2에 머물렀는데, 이는 특히 부적합항목인 유해물질의 기준·규격이 신설되어 해당 양식장에서 검사기록이 없는 경우와, 매우 드물게 발생하여 10년 동안 검출이력이 없어 학습데이터세트에는 없는 경우이었다. 베이지안 네트워크를 적용하여 부적합확률이 높은 생산단계 안전성 조사대상을 선정하게 되면 설명변수별로 시나리오에 따라 부적합확률을 설명가능하게 되어 다른 머신러닝 알고리즘을 적용하는 경우 지적되어온 설명불가능이라는 문제점을 해소할 수 있으며, 향후 안전성조사 데이터 축적 시 용이하게 모델 업데이트가 가능하며 이를 통해 모델의 예측성능개선도 기대할 수 있다는 장점이 있다.

• 한국물환경학회지
• /
• 제29권6호
• /
• pp.832-841
• /
• 2013

In recent years, eutrophication in the Paldang Lake has become one of the major environmental problems in Korea as it may threaten drinking water safety and human health. Thus it is important to understand the phenomena and predict the time and magnitude of algal blooms for applying adequate algal reduction measures. This study performed seasonal water quality assessment and chlorophyll-a prediction using Bayseian simple/multiple linear regression analysis. Bayseian regression analysis could be a useful tool to overcome limitations of conventional regression analysis. Also it can consider uncertainty in prediction by using posterior distribution. Generally, chlorophyll-a of a P2(Paldang Dam 2) site showed high concentration in spring and it was similar to that of P4(Paldang Dam 4) site. For the development of Bayseian model, we performed seasonal correlation. As a result, chlorophyll-a of a P2 site had a high correlation with P5(Paldang Dam 5) site in spring (r = 0.786, p<0.05) and with P4 in winter (r = 0.843, p<0.05). Based on the DIC (Deviance Information Criterion) value, critical explanatory variables of the best fitting Bayesian linear regression model were selected as a $PO_4-P$ (P2), Chlorophyll-a (P5) in spring, $NH_3-N$ (P2), Chlorophyll-a (P4), $NH_3-N$ (P4) in summer, DTP (P2), outflow (P2), TP (P3), TP (P4) fall, COD (P2), Chl-a (P4) and COD (P4) in winter. The results of chlorophyll-a prediction showed relatively high $R^2$ and low RMSE values in summer and winter.

• Animal Bioscience
• /
• 제37권8호
• /
• pp.1333-1344
• /
• 2024

Objective: Litter size and piglet loss at birth significantly impact piglet production and are closely associated with sow parity. Understanding how these traits vary across different parities is crucial for effective herd management. This study investigates the patterns of the number of born alive piglets (NBA), number of piglet losses (NPL), and the proportion of piglet losses (PPL) at birth in Landrace sows under tropical conditions. Additionally, it aims to identify the most suitable model for describing these patterns. Methods: A dataset comprising 2,322 consecutive reproductive records from 258 Landrace sows, spanning parities from 1 to 9, was analyzed. Modeling approaches including 2nd and 3rd degree polynomial models, the Wood gamma function, and a longitudinal model were applied at the individual level to predict NBA, NPL, and PPL. The choice of the best-fitting model was determined based on the lowest mean and standard deviation of the difference between predicted and actual values, Akaike information criterion (AIC), and Bayesian information criterion (BIC). Results: Sow parity significantly influenced NBA, NPL, and PPL (p<0.0001). NBA increased until the 4th parity and then declined. In contrast, NPL and PPL decreased until the 2nd parity and then steadily increased until the 8th parity. The 2nd and 3rd degree polynomials, and longitudinal models showed no significant differences in predicting NBA, NPL, and PPL (p>0.05). The 3rd degree polynomial model had the lowest prediction standard deviation and yielded the smallest AIC and BIC. Conclusion: The 3rd degree polynomial model offers the most suitable description of NBA, NPL, and PPL patterns. It holds promise for applications in genetic evaluations to enhance litter size and reduce piglet loss at birth in sows. These findings highlight the importance of accounting for sow parity effects in swine breeding programs, particularly in tropical conditions, to optimize piglet production and sow performance.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권4호
• /
• pp.1730-1747
• /
• 2018

Multi-view video plus depth (MVD) is a mainstream format of 3D scene representation in free viewpoint video systems. The advanced 3D extension of the high efficiency video coding (3D-HEVC) standard introduces new prediction tools to improve the coding performance of depth video. However, the depth video in 3D-HEVC is time consuming. To reduce the complexity of the depth video inter coding, we propose a fast coding unit (CU) size and mode decision algorithm. First, an off-line trained Bayesian model is built which the feature vector contains the depth levels of the corresponding spatial, temporal, and inter-component (texture-depth) neighboring largest CUs (LCUs). Then, the model is used to predict the depth level of the current LCU, and terminate the CU recursive splitting process. Finally, the CU mode search process is early terminated by making use of the mode correlation of spatial, inter-component (texture-depth), and inter-view neighboring CUs. Compared to the 3D-HEVC reference software HTM-10.0, the proposed algorithm reduces the encoding time of depth video and the total encoding time by 65.03% and 41.04% on average, respectively, with negligible quality degradation of the synthesized virtual view.

• 농업과학연구
• /
• 제39권3호
• /
• pp.357-364
• /
• 2012

Genomic breeding value (GEBV) has recently become available in the beef cattle industry. Genomic selection methods are exceptionally valuable for selecting traits, such as marbling, that are difficult to measure until later in life. One method to utilize information from sparse marker panels is the Bayesian model selection method with RJMCMC. The accuracy of prediction varies between a multiple SNP model with RJMCMC (0.47 to 0.73) and a least squares method (0.11 to 0.41) when using SNP information, while the accuracy of prediction increases in the multiple SNP (0.56 to 0.90) and least square methods (0.21 to 0.63) when including a polygenic effect. In the multiple SNP model with RJMCMC model selection method, the accuracy ($r^2$) of GEBV for marbling predicted based only on SNP effects was 0.47, while the $r^2$ of GEBV predicted by SNP plus polygenic effect was 0.56. The accuracies of GEBV predicted using only SNP information were 0.62, 0.68 and 0.73 for CWT, EMA and BF, respectively. However, when polygenic effects were included, the accuracies of GEBV were increased to 0.89, 0.90 and 0.89 for CWT, EMA and BF, respectively. Our data demonstrate that SNP information alone is missing genetic variation information that contributes to phenotypes for carcass traits, and that polygenic effects compensate genetic variation that whole genome SNP data do not explain. Overall, the multiple SNP model with the RJMCMC model selection method provides a better prediction of GEBV than does the least squares method (single marker regression).

• 한국정보시스템학회지:정보시스템연구
• /
• 제26권3호
• /
• pp.47-61
• /
• 2017

Purpose Although automobile production has increased since the development of the Korean automobile industry, the number of customers who can purchase automobiles decreases relatively. Therefore, automobile companies need to develop strategies to attract customers and promote their repurchase behaviors. To this end, this paper analyzed customer data from a Korean automobile company using data mining techniques to derive repurchase strategies. Design/methodology/approach We conducted under-sampling to balance the collected data and generated 10 datasets. We then implemented prediction models by applying a decision tree, naive Bayesian, and artificial neural network algorithms to each of the datasets. As a result, we derived 10 patterns consisting of 11 variables affecting customers' decisions about repurchases from the decision tree algorithm, which yielded the best accuracy. Using the derived patterns, we proposed helpful strategies for improving repurchase rates. Findings From the top 10 repurchase patterns, we found that 1) repurchases in January are associated with a specific residential region, 2) repurchases in spring or autumn are associated with whether it is a weekend or not, 3) repurchases in summer are associated with whether the automobile is equipped with a sunroof or not, and 4) a customized promotion for a specific occupation increases the number of repurchases.

• Journal of the Korean Data and Information Science Society
• /
• 제24권4호
• /
• pp.857-865
• /
• 2013

집중 호우로 인한 피해가 증가하면서 다양한 기법들을 이용하여 강우량 예측에 대한 관심이 높아졌다. 최근에는 극단분포를 활용하여 강우량을 예측하려는 시도가 늘고 있다. 본 연구에서는 일반화 극단 분포를 활용하여 실제 서울시의 1973년부터 2010년까지 7월달의 사후예측분포를 생성하고, 수치적인 계산을 위해서 MCMC (Markov chain Monte Carlo)알고리즘을 활용하였다. 이 연구를 통해서 사후예측분포의 점추정값들을 비교하였고 2011년 7월달의 자료와 비교해 봤을 때 집중 호우의 확률이 증가한 것을 알 수 있었다.