• Korean Journal of Construction Engineering and Management
• /
• v.19 no.3
• /
• pp.33-42
• /
• 2018

In the aftermath of an earthquake, seismic-damaged infrastructure systems loss estimation is the first step for the disaster response. However, lifeline systems' ability to supply service can be volatile by external factors such as disturbances of nearby facilities, and not by own physical issue. Thus, this research develops the bayesian model for probabilistic inference on common-cause and cascading failure of seismic-damaged lifeline systems. In addition, the authors present network robustness estimation metrics in the context of failure propagation. In order to quantify the functional loss and observe the effect of the mitigation plan, power and water supply system in Daegu-Gyeongbuk in South Korea is selected as case network. The simulation results show that reduction of cascading failure probability allows withstanding the external disruptions from a perspective of the robustness improvement. This research enhances the comprehensive understanding of how a single failure propagates to whole lifeline system performance and affected region after an earthquake.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.4B
• /
• pp.354-361
• /
• 2011

Increasing interest and technological advances in smart home has led to active research on context-awareness service and prediction algorithms such as Bayesian Networks, Tree-Dimensional Structures and Genetic prediction algorithms. Context-awareness service presents that providing automatic customized service regarding individual user's pattern surely helps users improve the quality of life. However, it is difficult to implement context-awareness service because the problems are that handling coincidence with context information and exceptional cases have to consider. To overcome this problem, we proposes an Intelligent Sequential Matching Algorithm(ISMA), models context-awareness service using Timed Petri-net(TPN) which is petri-net to have time factor. The example scenario illustrates the effectiveness of the Timed Petri-net model and our proposed algorithm improves average 4~6% than traditional in the accuracy and reliability of prediction.

• Journal of KIISE
• /
• v.43 no.7
• /
• pp.795-800
• /
• 2016

In image classification, recent CNNs compete with human performance. However, there are limitations in more general recognition. Herein we deal with indoor images that contain too much information to be directly processed and require information reduction before recognition. To reduce the amount of data processing, typically variational inference or variational Bayesian methods are suggested for object detection. However, these methods suffer from the difficulty of marginalizing over the given space. In this study, we propose an image-text integrated recognition system using active vision based on Spatial Transformer Networks. The system attempts to efficiently sample a partial region of a given image for a given language information. Our experimental results demonstrate a significant improvement over traditional approaches. We also discuss the results of qualitative analysis of sampled images, model characteristics, and its limitations.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.8
• /
• pp.3984-4005
• /
• 2017

Image recognition has become an increasingly important topic for its wide application. It is highly challenging when facing to large-scale database with large variance. The recognition systems rely on a key component, i.e. the low-level feature or the learned mid-level feature. The recognition performance can be potentially improved if the data distribution information is exploited using a more sophisticated way, which usually a function over hidden variable, model parameter and observed data. These methods are called generative score space. In this paper, we propose a discriminative extension for the existing generative score space methods, which exploits class label when deriving score functions for image recognition task. Specifically, we first extend the regular generative models to class conditional models over both observed variable and class label. Then, we derive the mid-level feature mapping from the extended models. At last, the derived feature mapping is embedded into a discriminative classifier for image recognition. The advantages of our proposed approach are two folds. First, the resulted methods take simple and intuitive forms which are weighted versions of existing methods, benefitting from the Bayesian inference of class label. Second, the probabilistic generative modeling allows us to exploit hidden information and is well adapt to data distribution. To validate the effectiveness of the proposed method, we cooperate our discriminative extension with three generative models for image recognition task. The experimental results validate the effectiveness of our proposed approach.

• Experimental and Molecular Medicine
• /
• v.50 no.11
• /
• pp.7.1-7.12
• /
• 2018

Recent findings from The Cancer Genome Atlas project have provided a comprehensive map of genomic alterations that occur in hepatocellular carcinoma (HCC), including unexpected mutations in apolipoprotein B (APOB). We aimed to determine the clinical significance of this non-oncogenetic mutation in HCC. An Apob gene signature was derived from genes that differed between control mice and mice treated with siRNA specific for Apob (1.5-fold difference; P < 0.005). Human gene expression data were collected from four independent HCC cohorts (n = 941). A prediction model was constructed using Bayesian compound covariate prediction, and the robustness of the APOB gene signature was validated in HCC cohorts. The correlation of the APOB signature with previously validated gene signatures was performed, and network analysis was conducted using ingenuity pathway analysis. APOB inactivation was associated with poor prognosis when the APOB gene signature was applied in all human HCC cohorts. Poor prognosis with APOB inactivation was consistently observed through cross-validation with previously reported gene signatures (NCIP A, HS, high-recurrence SNUR, and high RS subtypes). Knowledge-based gene network analysis using genes that differed between low-APOB and high-APOB groups in all four cohorts revealed that low-APOB activity was associated with upregulation of oncogenic and metastatic regulators, such as HGF, MTIF, ERBB2, FOXM1, and CD44, and inhibition of tumor suppressors, such as TP53 and PTEN. In conclusion, APOB inactivation is associated with poor outcome in patients with HCC, and APOB may play a role in regulating multiple genes involved in HCC development.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2015.05a
• /
• pp.22-22
• /
• 2015

가뭄은 적시에 경보해야 하는 홍수와 달리 진행속도가 느리고 시간적으로 대처할 여유가 있어 진행중일지라도 미리 감지만 한다면 그 피해를 최소화할 수 있다. 이로 인해 미국 등 수문기상 선진국에서는 수문기상 장기예보자료로부터 가뭄전망정보 생산기술을 개발하였으며, 특히 가뭄전망의 정확도 향상을 위해 여러 통계적 보정기법을 적용하고 있다. 국내의 경우 기상청에서 가뭄전망을 목적으로 2011년에 수치예보모델을 이용하여 가뭄전망정보를 생산한바 있으나, 전망정보의 불확실성 문제로 가뭄예보에 활용하는데 한계가 있어 이를 개선할 수 있는 기술개발이 요구되는 실정이다. 본 연구에서는 기후예측자료를 이용하여 가뭄전망정보 생산기술을 개발하고 정확도 개선을 위해 베이지안 기법을 연계하였다. GloSea5 (Global Seasonal forecast model 5) 장기예보자료를 이용하였으며, 베이지안 기법을 통해 과거 관측자료에 대한 사전분포, 모델의 전망정보로부터 우도함수를 유도하여 최종 사후분포를 추정하였다. 베이지안 기법 적용 전 후에 따른 가뭄지수를 산정하였다. 관측자료 기반의 가뭄지수와의 비교분석을 통해 선행기간 및 계절별 가뭄예측 성능을 평가하였으며, 실제 가뭄기간 동안에 가뭄의 재현성을 지역별로 분석하였다. 장기예보자료만을 활용한 기존 가뭄전망에서는 관측 자료가 포함된 1개월 전망에서도 불확실성이 매우 높았지만 베이지안 기법 적용으로 가뭄전망의 정확도가 크게 개선되었다. 특히, 1, 2개월 전망의 시계열 가뭄지수가 관측기반의 가뭄지수의 거동과 매우 유사하게 나타났으며, 지역별로도 베이지안 기법 적용시 실제 가뭄피해 상황을 적절히 재현하는 것으로 나타났다. 국내 가뭄예보에 있어 기후예측정보를 단순활용하기 보다는 베이지안과 같은 통계적 보정기법을 이용하여 가뭄전망정보를 생산하는 것이 바람직하며, 본 연구에서는 가뭄예보업무에 활용될 수 있도록 베이지안 기법에 대한 검증 및 평가를 지속적으로 수행할 계획이다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2019.05a
• /
• pp.100-100
• /
• 2019

최근 대기 중 미세먼지의 농도가 높은 일수가 급증하면서, 미세먼지를 저감하고자 하는 연구가 활발히 이루어지고 있다. 미세먼지는 주로 자동차 혹은 공장 등 인간 활동에 의한 오염물질 배출에 의해 발생하는 것으로 알려져 있으며, 태양복사에너지, 토양수분, 강우, 풍속 등의 수문기상학적 인자에 의해 발생, 이동, 소멸의 과정을 거친다. 현재 우리나라에서는 미세먼지 농도를 관측하기 위해 지점 기반의 관측소를 운영하고 있으며, 관측소가 위치하지 않은 지역의 미세먼지 농도는 선형 보간법 등을 활용한 내삽 기법을 통해 제공하고 있다. 그러나 미세먼지 농도는 다양한 수문기상인자들의 영향에 의한 차이가 크게 나타나기 때문에 지점 기반의 자료로는 해당 지역의 미세먼지 농도를 추정하는 데 어려움이 많다. 본 연구에서는 미세먼지의 공간적인 분포를 추정하고자 MODerate resolution Imaging Spectroradiometer (MODIS) 에어로졸 자료와 Global Land Data Assimilation System (GLDAS) 수문기상인자를 활용하여 미세먼지 농도에 영향을 주는 것으로 판단되는 다양한 수문기상인자들과의 상관성을 분석하였다. 미세먼지와 각 인자간의 상관성을 분석하여 높은 상관성을 갖는 수문기상인자들을 도출하고 최적의 선형회귀분석 모델을 구축하기 위해 베이지안 모델 평균(Bayesian Model Averaging, BMA)을 사용하였으며, 지점 데이터와의 비교를 통해 활용성을 검증하였다. 전체적으로 수문기상인자를 사용한 선형회귀분석 결과에서는 미세먼지농도 변화의 경향을 반영하고 있는 것을 확인할 수 있었으나, 계절별, 지역별 등 대기 특성을 고려하지 않아 각 기간의 급격한 농도 변화를 감지하기에 어려움이 있었다. 이러한 연구를 바탕으로 수문기상인자와 미세먼지 농도의 패턴이 더욱 정확히 분석된다면, 미세먼지 농도 모니터링과 정확한 예보 시스템의 구축에 효과적으로 활용 될 것으로 기대된다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2017.05a
• /
• pp.411-411
• /
• 2017

유역 내 발생하는 강우의 공간적인 분포는 인접성 및 거리에 따라 달라질 수 있다. 공간자기상관 분석은 공간단위(유역 또는 행정구역)의 변수(강수 등)가 주변지역과 갖는 관계를 통해 얼마나 분산되어 있는지 혹은 군집되어 있는지를 판별하는 기법으로 최근 많은 연구에서 활성화 되고 있다. 본 연구에서는 낙동강유역을 대상으로 1980~2000년까지 20개년의 기상청을 통해 수집한 강우자료와 CMIP5(Coupled Model Intercomparison Project Phase 5)에서 제공하는 기후변화 자료 중 가용할 수 있는 20개 모델의 강우를 수집하였다. 기후변화 자료는 정상성 분위사상법으로 지역오차보정을 실시하고 불확실성을 저감하고자 베이지안 모델 평균기법을 통해 새로운 시계열을 생성하였다. 생성된 시계열의 공간적인 분포를 정량적으로 평가하고자 중권역별 공간자기상관 분석을 수행하였다. 대부분의 연구에서는 GIS를 활용하여 정성적으로 강우의 분포를 나타내고 있지만 본 연구에서는 공간단위의 인접성 또는 거리에 따른 척도를 기반으로 공간자기상관을 탐색할 수 있는 Moran's I와 LISA(Local Indicators of Spatial Association)기법을 적용하였다. Moran's I는 전체 연구지역에 대한 관계를 하나의 값으로 보여주는 전역적인 기법이며, LISA는 상대적으로 넓은 지역을 국지적으로 구분하여 특정지역에 대한 Hot spot 및 Cold spot을 통해 공간자기상관 정도를 나타내는 국지적인 기법이다. 두 기법을 적용하기 위하여 인접성 기반의 공간매트릭스를 산정하고 계절별 관측값과 베이지안 앙상블 강우의 Moran's I 및 LISA 분석을 실시하였다. 관측자료와 베이지안 앙상블 강우의 분석결과가 매우 유사하게 나타남으로써 베이지안 앙상블 강우의 공간적인 분포가 관측강우를 충분히 재현하고 있다고 판단된다.

• Animal Bioscience
• /
• v.34 no.6
• /
• pp.967-974
• /
• 2021

Objective: The objective of this study was to estimate the genetic correlation (rpc) of growth performance between purebred (Duroc and Korean native) and synthetic (WooriHeukDon) pigs using a single-step method. Methods: Phenotypes of 15,902 pigs with genotyped data from 1,792 pigs from a nucleus farm were used for this study. We estimated the rpc of several performance traits between WooriHeukDon and purebred pigs: day of target weight (DAY), backfat thickness (BF), feed conversion rate (FCR), and residual feed intake (RFI). The variances and covariances of the studied traits were estimated by an animal multi-trait model that applied the Bayesian inference. Results: rpc within traits was lower than 0.1 for DAY and BF, but high for FCR and RFI; in particular, rpc for RFI between Duroc and WooriHeukDon pigs was nearly 1. Comparison between different traits revealed that RFI in Duroc pigs was associated with different traits in WooriHeukDon pigs. However, the most of rpc between different traits were estimated with low or with high standard deviation. Conclusion: The results indicated that there were substantial differences in rpc of traits in the synthetic WooriHeukDon pigs, which could be caused by these pigs having a more complex origin than other crossbred pigs. RFI was strongly correlated between Duroc and WooriHeukDon pigs, and these breeds might have similar single nucleotide polymorphism effects that control RFI. RFI is more essential for metabolism than other growth traits and these metabolic characteristics in purebred pigs, such as nutrient utilization, could significantly affect those in synthetic pigs. The findings of this study can be used to elucidate the genetic architecture of crossbred pigs and help develop new breeds with target traits.

• Horticulture, Environment, and Biotechnology : HEB
• /
• v.59 no.6
• /
• pp.875-888
• /
• 2018

Tulip (Tulipa L.) is one of the most important ornamental geophytes in the world. Analysis of molecular variability of tulips is of great importance in conservation and parental lines selection in breeding programs. Of the 70 genic microsatellites, 15 highly polymorphic and reproducible markers were used to assess the genetic diversity, structure, and relationships among 280 individuals of 36 wild and cultivated tulip accessions from two countries: Iran and the Netherlands. The mean values of gene diversity and polymorphism information content were 0.69 and 0.66, respectively, which indicated the high discriminatory power of markers. The calculated genetic diversity parameters were found to be the highest in wild T. systola Stapf (Derak region). Bayesian model-based STRU CTU RE analysis detected five gene pools for 36 germplasms which corresponded with morphological observations and traditional classifications. Based on analysis of molecular variance, to conserve wild genetic resources in some geographical locations, sampling should be performed from distant locations to achieve high diversity. The unweighted pair group method with arithmetic mean dendrogram and principal component analysis plot indicated that among wild tulips, T. systola and T. micheliana Hoog exhibited the closest relationships with cultivated tulips. Thus, it can be assumed that wild tulips from Iran and perhaps other Middle East countries played a role in the origin of T. gesneriana, which is likely a tulip species hybrid of unclear origin. In conclusion, due to the high genetic variability of wild tulips, they can be used in tulip breeding programs as a source of useful alleles related to resistance against stresses.