• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.5
• /
• pp.684-690
• /
• 2007

We describe our method to predict the direction of conditional probabilistic dependencies between clusters of random variables. Selected variables called 'gateway variables' are used to predict the conditional probabilistic dependency relations between clusters. The direction of conditional probabilistic dependencies between clusters are predicted by finding directed acyclic graph (DAG)-shaped dependency structure between the gateway variables. We show that our method shows meaningful prediction results in determining directions of conditional probabilistic dependencies between clusters.

• Nuclear Engineering and Technology
• /
• v.54 no.1
• /
• pp.110-116
• /
• 2022

Human failure event (HFE) dependency analysis is a part of human reliability analysis (HRA). For efficient HFE dependency analysis, a maximum number of minimal cut sets (MCSs) that have HFE combinations are generated from the fault trees for the probabilistic safety assessment (PSA) of nuclear power plants (NPPs). After collecting potential HFE combinations, dependency levels of subsequent HFEs on the preceding HFEs in each MCS are analyzed and assigned as conditional probabilities. Then, HFE recovery is performed to reflect these conditional probabilities in MCSs by modifying MCSs. Inappropriate HFE dependency analysis and HFE recovery might lead to an inaccurate core damage frequency (CDF). Using the above process, HFE recovery is performed on MCSs that are generated with a non-zero truncation limit, where many MCSs that have HFE combinations are truncated. As a result, the resultant CDF might be underestimated. In this paper, a new method is suggested to incorporate HFE recovery into the MCS generation stage. Compared to the current approach with a separate HFE recovery after MCS generation, this new method can (1) reduce the total time and burden for MCS generation and HFE recovery, (2) prevent the truncation of MCSs that have dependent HFEs, and (3) avoid CDF underestimation. This new method is a simple but very effective means of performing MCS generation and HFE recovery simultaneously and improving CDF accuracy. The effectiveness and strength of the new method are clearly demonstrated and discussed with fault trees and HFE combinations that have joint probabilities.

• The Korean Journal of Applied Statistics
• /
• v.32 no.5
• /
• pp.763-782
• /
• 2019

This paper proposes a probabilistic document ranking model incorporating term dependencies. Document ranking is a fundamental information retrieval task. The task is to sort documents in a collection according to the relevance to the user query (Qin et al., Information Retrieval Journal, 13, 346-374, 2010). A probabilistic model is a model for computing the conditional probability of the relevance of each document given query. Most of the widely used models assume the term independence because it is challenging to compute the joint probabilities of multiple terms. Words in natural language texts are obviously highly correlated. In this paper, we assume a multinomial distribution model to calculate the relevance probability of a document by considering the dependency structure of words, and propose an information retrieval model to rank a document by estimating the probability with the maximum entropy method. The results of the ranking simulation experiment in various multinomial situations show better retrieval results than a model that assumes the independence of words. The results of document ranking experiments using real-world datasets LETOR OHSUMED also show better retrieval results.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07b
• /
• pp.727-729
• /
• 2005

나이브베이즈분류기($na\ddot{i}ve$ Bayes classifier)는 학습, 적용 및 계산자원 이용의 측면에서 매우 효율적인 모델이다. 또한, 그 분류 성능 역시 다른 기법에 비해 크게 떨어지지 않음이 다양한 실험을 통해 보여져 왔다. 특히, 데이터를 생성한 실제 확률분포를 나이브베이즈분류기가 정확하게 표현할 수 있는 경우에는 최대의 효과를 볼 수 있다. 하지만, 실제 확률분포에 존재하는 조건부독립성(conditional independence)이 나이브베이즈분류기의 구조와 일치하지 않는 경우에는 성능이 하락할 수 있다. 보다 구체적으로, 각 자질변수(feature variable)들 사이에 확률적 의존관계(probabilistic dependency)가 존재하는 경우 성능 하락은 심화된다. 본 논문에서는 이러한 나이브베이즈분류기의 약점을 효율적으로 해결할 수 있는 자질변수의 통합기법을 제시한다. 자질변수의 통합은 각 변수들 사이의 관계를 명시적으로 표현해 주는 방법이며, 특히 상호정보량(mutual information)에 기반한 통합 변수의 선정이 성능 향상에 크게 기여함을 실험을 통해 보인다.

• Journal of the Korean Society of Safety
• /
• v.38 no.3
• /
• pp.110-120
• /
• 2023

Following the Fukushima accident, portable equipment employed as accident mitigating systems have been installed and operated to reduce core damage and large early release frequencies. In addition, the establishment of an accident management strategy has gained importance. This study investigated the current status of portable equipment including the international portable equipment FLEX (diverse and flexible coping strategies), and domestic portable equipment multi-barrier accident coping strategy (MACST). Research on optimal utilization of MACST remains insufficient. As a preliminary study for establishing an optimal strategy, sensitivity studies were conducted to facilitate the priority of use on portable equipment, number of portable equipment, and dependency of operator actions based on a multi-unit probabilistic safety assessment model. The results revealed the conditions that reduced the multi-unit and site conditional core damage probabilities, indicating the optimal strategy of MACST. The results of this study can be used as a reference for establishing an optimal strategy that utilizes domestic safety equipment in the future.

• Nuclear Engineering and Technology
• /
• v.54 no.12
• /
• pp.4601-4619
• /
• 2022

The classical Probabilistic Safety Analysis (PSA) does not include any time dependence explicitly. However, the success criteria (SC) could evolve during the cycle for some initiating events. In that sense, there is a type of sequence in which this time-dependency is quite important, the family of Anticipated Transient without Scram (ATWS) sequences in Pressurized Water Reactors. Therefore, a new risk-informed approach is proposed in this paper, which makes it possible to obtain the time-dependent SC evolution of the safety functions affected by the Moderator Temperature Coefficient (MTC) value. Then, the evolution of the ATWS conditional core damage probability (CCDP) could be obtained using a PSA model. To quantify the CCDP, the average values of the time-dependent failure probabilities must be computed. Finally, the comparison between the CCDP obtained through the application of the classical PSA approach and the new one makes it possible to quantify the impact of time-dependence on the SC of the headers that this new risk-informed ATWS approach can provide.