• 한국ITS학회:학술대회논문집
• /
• v.2006 no.10
• /
• pp.286-289
• /
• 2006

• The KIPS Transactions:PartB
• /
• v.15B no.2
• /
• pp.147-158
• /
• 2008

Fuzzy Cognitive Map (FCM) and Bayesian Belief Network (BBN) are two major frameworks for modeling, representing and reasoning about causal knowledge. Despite their extensive use in causal knowledge engineering, there is no reported work which compares their respective roles. This paper aims to fill the gap by providing a qualitative comparison of the two frameworks through a systematic analysis based on some inherent features of the frameworks. We proposed a set of comparison criteria which covers the entire process of causal knowledge engineering, including modeling, representation, and reasoning. These criteria are usability, expressiveness, reasoning capability, formality, and soundness. The results of comparison have revealed some important facts about the characteristics of FCM and BBN, which will help to determine how FCM and BBN should be used, with respect to each other, in causal knowledge engineering.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.26 no.5
• /
• pp.335-342
• /
• 2016

When dynamics changes occurred in an existing Bayesian Network (BN), the related parameters embedding on the BN have to be updated to new parameters adapting to changed patterns. In this case, these parameters have to be updated with the consideration of the causalities in the BN. This research suggests a framework for updating parameters dynamically using Expectation Maximization (EM) algorithm and Harmony Search (HS) algorithm among several Meta-Heuristics techniques. While EM is an effective algorithm for estimating hidden parameters, it has a limitation that the generated solution converges a local optimum in usual. In order to overcome the limitation, this paper applies HS for tracking the global optimum values of Maximum Likelihood Estimators (MLE) of parameters. The proposed method suggests a learning and propagation framework of BN with dynamic changes for overcoming disadvantages of EM algorithm and converging a global optimum value of MLE of parameters.

• The Korean Journal of Applied Statistics
• /
• v.33 no.3
• /
• pp.309-320
• /
• 2020

Bayesian networks, also known as directed acyclic graphs (DAG), are used in many areas of medicine, meteorology, and genetics because relationships between variables can be modeled with graphs and probabilities. In particular, Bayesian network classifiers, which are used to predict discrete data, have recently become a new method of data mining. Bayesian networks can be grouped into different models that depend on structured learning methods. In this study, Bayesian network models are learned with various properties of structure learning. The models are compared to the simplest method, the naïve Bayes model. Classification results are compared by applying learned models to various real data. This study also compares the relationships between variables in the data through graphs that appear in each model.

• Journal of KIISE:Computing Practices and Letters
• /
• v.12 no.1
• /
• pp.56-65
• /
• 2006

Recently the study on service robots has been proliferated in many fields, and there are active developments for indoor services such as supporting for elderly people. It is important for robot to recognize objects and situations appropriately for effective and accurate service. Conventional object recognition methods have been based on the pre-defined geometric models, but they have limitations in indoor environments with uncertain situation such as the target objects are occluded by other ones. In this paper we propose a Bayesian network model to reason the probability of target objects for effective detection. We model the relationships between objects by activities, which are applied to non-static environments more flexibly. Overall structure is constructed by combining common-cause structures which are the units making relationship between objects, and it makes design process more efficient. We test the performance of two Bayesian networks for verifying the proposed Bayesian network model through experiments, resulting in accuracy of $86.5\%$ and $89.6\%$ respectively.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07b
• /
• pp.634-636
• /
• 2005

지능형 로봇 기술이 발전하면서 영상에서 장면을 이해하는 연구가 많은 관심을 받고 있으며, 최근에는 불확실한 환경에서도 좋은 성능을 발휘할 수 있는 확률적 접근 방법이 많이 연구되고 있다. 본 논문에서는 확률적 모델링이 가능한 베이지안 네트워크(BN)를 이용해서 장면 인식 추론 모듈을 설계하고, 실제 환경에서 얻어진 증거 및 베이지안 추론 결과를 결합하여 분류 성능을 향상시키기 위한 방법을 제안한다. 영상 정보는 시간에 대해 연속성을 가지고 있기 때문에, 증거 정보와 베이지안 추론 결과들을 적절히 결합하면 더 좋은 결과를 예상할 수 있으며, 본 논문에서는 확신 요소(Certainty Factor: CF) 분석에 의한 결합 방법을 사용하였다. 성능 평가 실험을 위해서 SET (Scale Invariant Feature Transform) 기법을 이용하여 물체 인식 처리를 수행하고, 여기서 얻어진 데이터를 베이지안 추론의 증거로 사용하였으며, 전문가의 CF 값 정의에 의한 베이지안 네트워크 설계 방법을 이용하였다.

• Journal of KIISE:Computing Practices and Letters
• /
• v.10 no.1
• /
• pp.89-98
• /
• 2004

Conversational agent is a system that provides users with proper information and maintains the context of dialogue on the natural language. Analyzing and modeling process of user's query is essential to make it more realistic, for which Bayesian network is a promising technique. When experts design the network for a domain, the network is usually very complicated and is hard to be understood. The separation of variables in the domain reduces the size of networks and makes it easy to design the conversational agent. Composing Bayesian network as two stages, we aim to design conversational agent easily and analyze user's query in detail. Also, previous information of dialogue makes it possible to maintain the context of conversation. Actually implementing it for a guide of web pages, we can confirm the usefulness of the proposed architecture for conversational agent.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06b
• /
• pp.121-123
• /
• 2006

베이지안 네트워크 병합은 여러 개의 베이지안 네트워크를 하나의 네트워크로 합치는 것을 말한다. 일반적으로 사용되는 병합 알고리즘은 병합 순서에 따라 최종결과 네트워크의 복잡도가 달라지는 문제를 갖고 있고, 최종 병합 네트워크의 에지 수를 최소화하는 병합 순서를 찾는 것은 NP-hard라고 증명되었다. 본 논문에서는 최적의 병합 순서를 결정하기 위해 진화 알고리즘을 사용하는 방법을 제안한다. 해공간 분석을 통해 permutation index 표현방법이 해탐색에 유리함을 보이고 이를 이용한 진화 알고리즘을 제안한다. 실험결과, 기존의 휴리스틱과 greedy 탐색 방법에 비해 제안한 방법이 우수한 성능을 보였다.

• Journal of KIISE:Software and Applications
• /
• v.31 no.12
• /
• pp.1569-1580
• /
• 2004

One commonly used approach to deal with uncertainty is Bayesian network which represents joint probability distributions of domain. There are some attempts to team the structure of Bayesian networks automatically and recently many researchers design structures of Bayesian network using evolutionary algorithm. However, most of them use the only one fittest solution in the last generation. Because it is difficult to combine all the important factors into a single evaluation function, the best solution is often biased and less adaptive. In this paper, we present a method of generating diverse Bayesian network structures through fitness sharing and combining them by Bayesian method for adaptive inference. In order to evaluate performance, we conduct experiments on learning Bayesian networks with artificially generated data from ASIA and ALARM networks. According to the experiments with diverse conditions, the proposed method provides with better robustness and adaptation for handling uncertainty.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2014.11a
• /
• pp.865-867
• /
• 2014

이 논문에서는 초등학생의 위험 상황을 자동으로 판단하기 위한 행위네트워크 설계 기법을 제안한다. 초등학생이 가지고 있는 스마트폰 센서로 주변 상황을 측정하고 측정한 값은 행위네트워크에 적용한다. 행위네트워크는 자동으로 위험상황을 인지하기 위해서 베이지안 확률을 도입한다. 반복적인 위험 상황을 학습함으로써 긴급시 빠르게 대처할 수 있다.