• 한국인터넷방송통신학회논문지
• /
• 제13권5호
• /
• pp.37-47
• /
• 2013

텍스트 마이닝은 비정형화된 텍스트를 분석하여 그 안에 내재된 패턴, 추세, 분포 등의 고급정보들을 추출하는 분야이다. 텍스트 마이닝은 기본적으로 비정형 데이터를 가정하므로 텍스트를 단순화된 모델로 표현하는 것이 필요하다. 현재까지 가장 많이 사용되고 있는 모델은 텍스트를 단순한 단어들의 집합으로 표현한 벡터공간 모델이다. 그러나 최근 들어 단어들의 의미적 관계까지 표현하기 위해 그래프를 이용한 텍스트 표현 모델을 많이 사용하고 있다. 본 논문에서는 텍스트 마이닝을 위한 기존의 연구 중에서 그래프에 기반한 텍스트 표현 모델의 방법들과 그들의 특징들을 기술한다. 또한 그래프 기반 텍스트 마이닝의 향후 발전방향에 대해서도 논한다.

• 로봇학회논문지
• /
• 제19권2호
• /
• pp.196-202
• /
• 2024

The ability to understand given environments and plan a sequence of actions leading to goal state is crucial for personal service robots. With recent advancements in deep learning, numerous studies have proposed methods for state representation in planning. However, previous works lack explicit information about relationships between objects when the state observation is converted to a single visual embedding containing all state information. In this paper, we introduce graph-based state representation that incorporates both object and relationship features. To leverage these advantages in addressing the task planning problem, we propose a Graph Neural Network (GNN)-based subgoal prediction model. This model can extract rich information about object and their interconnected relationships from given state graph. Moreover, a search-based algorithm is integrated with pre-trained subgoal prediction model and state transition module to explore diverse states and find proper sequence of subgoals. The proposed method is trained with synthetic task dataset collected in simulation environment, demonstrating a higher success rate with fewer additional searches compared to baseline methods.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2018년도 추계학술발표대회
• /
• pp.984-987
• /
• 2018

In an Internet of Thing (IoT) environment, entities with different attributes and capacities are going to be connected in a highly connected fashion. Specifically, not only the mechanical and electronic devices but also other entities such as people, locations and applications are connected to each other. Understanding and managing these connections play an important role for businesses, which identify opportunities for new IoT services. Traditional approach for storing and querying IoT data is used of a relational database management system (RDMS) such as MySQL or MSSQL. However, using RDMS is not flexible and sufficient for handling heterogeneous IoT data because these data have deeply complex relationships which require nested queries and complex joins on multiple tables. In this paper, we propose a graph model for constructing a graph database of heterogeneous IoT data. Graph databases are purposely-built to store highly connected data with nodes representing entities and edges representing the relationships between these entities. Our model fuses social graph, spatial graph, and things graph, and incorporates the relationships among them. We then present a case study which applies our model for representing data from a Smart Campus using Neo4J platform. Through the results of querying to answer real questions in Smart Campus management, we show the viability of our model.

• 한국국방경영분석학회지
• /
• 제16권2호
• /
• pp.56-74
• /
• 1990

Programming-in-the-Large refers to software development environment and includes the organization and representation of a system structure, module decomposition, component dependence analysis, seperate compilation, subsystem and composition identification. The most intricate problem in this environment is the mastery of the structural complexity of large software systems. Ada programming language is tailored to the needs for building of large, integrated software systems from many program units. The visibility graph generating model presented in this paper transforms Ada source program into a visibility graph with nodes for program units and edges for visibility relations among program units. The system description in terms of program units and their visibility relations produced by this model can be utilized for some apects of Programming-in-the-Large environment and also assists designeers, programmers, integrators and maintainers in defining, understanding and exploring the structure of evolving software systems. The model designed and implemented in Ada programming language runs on PCs and will remain useful both in practice and as experimental tool.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권9호
• /
• pp.3598-3614
• /
• 2020

With the increase of motor vehicles and tourism demand, some traffic problems gradually appear, such as traffic congestion, safety accidents and insufficient allocation of traffic resources. Facing these challenges, a model of Spatio-Temporal Dilated Convolutional Network (STDGCN) is proposed for assistance of extracting highly nonlinear and complex characteristics to accurately predict the future traffic flow. In particular, we model the traffic as undirected graphs, on which graph convolutions are built to extract spatial feature informations. Furthermore, a dilated convolution is deployed into graph convolution for capturing multi-scale contextual messages. The proposed STDGCN integrates the dilated convolution into the graph convolution, which realizes the extraction of the spatial and temporal characteristics of traffic flow data, as well as features of road occupancy. To observe the performance of the proposed model, we compare with it with four rivals. We also employ four indicators for evaluation. The experimental results show STDGCN's effectiveness. The prediction accuracy is improved by 17% in comparison with the traditional prediction methods on various real-world traffic datasets.

• 대한안전경영과학회지
• /
• 제17권4호
• /
• pp.335-342
• /
• 2015

The increasing interests on patents have led many individuals and companies to apply for many patents in various areas. Applied patents are stored in the forms of electronic documents. The search and categorization for these documents are issues of major fields in data mining. Especially, the keyword extraction by which we retrieve the representative keywords is important. Most of techniques for it is based on vector space model. But this model is simply based on frequency of terms in documents, gives them weights based on their frequency and selects the keywords according to the order of weights. However, this model has the limit that it cannot reflect the relations between keywords. This paper proposes the advanced way to extract the more representative keywords by overcoming this limit. In this way, the proposed model firstly prepares the candidate set using the vector model, then makes the graph which represents the relation in the pair of candidate keywords in the set and selects the keywords based on this relationship graph.

• 한국차세대컴퓨팅학회논문지
• /
• 제13권2호
• /
• pp.102-110
• /
• 2017

최근 자연어 처리 분야에서 단어의 모호성을 해소하기 위해서 다양한 기계 학습 방법이 적용되고 있다. 지도 학습에 사용되는 데이터는 정답을 부착하기 위해 많은 비용과 시간이 필요하므로 최근 연구들은 비지도 학습의 성능을 높이기 위한 노력을 지속적으로 시도하고 있다. 단어 모호성 해소(word sense disambiguation)를 위한 비지도 학습연구는 지식 기반(knowledge base)를 이용한 방법들이 주목받고 있다. 이 방법은 학습 데이터 없이 지식 기반의 정보을 이용하여 문장 내에서 모호성을 가지는 단어의 의미를 결정한다. 지식 기반을 이용한 방법에는 그래프 기반방식과 유사도 기반 방법이 대표적이다. 그래프 기반 방법은 모호성을 가지는 단어와 그 단어가 가지는 다양한 의미들의 집합 간의 모든 경로에 대한 의미 그래프를 구축한다는 장점이 있지만 불필요한 의미 경로가 추가되어 오류를 증가시킨다는 단점이 있다. 이러한 문제를 해결하기 위해 본 논문에서는 그래프 구축을 위해 불필요한 간선들을 배제하면서 반복적으로 그래프를 재구축하는 모델을 제안한다. 또한, 구축된 의미 그래프에서 더욱 정확한 의미를 예측하기 위해 하이브리드 유사도 예측 모델을 적용한다. 또한 제안된 모델은 다국어 어휘 의미망 사전인 BabelNet을 사용하기 때문에 특정 언어뿐만 아니라 다양한 언어에도 적용 가능하다.

• Nuclear Engineering and Technology
• /
• 제29권6호
• /
• pp.433-443
• /
• 1997

In this work, an analytic approach to the dependability of software in the operational phase is suggested with special attention to the hardware fault effects on the software behavior : The hardware faults considered are memory faults and the dependability measure in question is the reliability. The model is based on the simple reliability theory and the graph theory which represents the software with graph composed of nodes and arcs. Through proper transformation, the graph can be reduced to a simple two-node graph and the software reliability is derived from this graph. Using this model, we predict the reliability of an application software in the digital system (ILS) in the nuclear power plant and show the sensitivity of the software reliability to the major physical parameters which affect the software failure in the normal operation phase. We also found that the effects of the hardware faults on the software failure should be considered for predicting the software dependability accurately in operation phase, especially for the software which is executed frequently. This modeling method is particularly attractive for the medium size programs such as the microprocessor-based nuclear safety logic program.

• 동의생리병리학회지
• /
• 제26권6호
• /
• pp.818-822
• /
• 2012

We in this paper propose a prognosis graph, analyzing prognoses of each pattern described in the Korean medicine literatures. This graph is represented as the integrated graphs about knowledge of patterns and their transitions in the prognoses, where a node becomes a pattern name and a edge becomes a transition between patterns, along with a condition with respect to cause or mechanism of the pattern. The knowledge of prognoses which a pattern is transit into another pattern can be identified at a glance by using this model. We also construct a upper-level prognosis graph, excluding five viscera and six entrails from the model. This upper-level prognosis graph contains the conceptual knowledge than clinical one so that it may be helpful to students and researchers in the Korean medicine fields.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제9권8호
• /
• pp.243-250
• /
• 2020

지식 그래프 기반의 질문 응답 문제는 자연어 질문들에 대한 깊은 이해뿐만 아니라, 대규모 지식 그래프 상에서 올바른 답변을 찾기 위한 효과적인 추론 능력을 필요로 한다. 본 논문에서는 다중 홉 추론을 요구하는 복잡한 자연어 질문에 대해 연관 지식 그래프 위에서 답변 추론을 효과적으로 수행할 수 있는 심층 신경망 모델을 제안한다. 제안 모델에서는 지식 그래프 상의 각 개체 노드와 이웃 노드 간의 양방향 특징 전파를 허용할뿐만 아니라, 두 이웃 노드 쌍 간의 맥락 정보까지 활용할 수 있는, 표현력이 뛰어난 쌍 선형 그래프 신경망(BGNN)을 이용한다. 본 논문에서는 오픈 도메인의 지식 베이스인 Freebase, 자연어 질문 응답을 위한 벤치마크 데이터 집합들인 WebQuestionsSP와 MetaQA를 이용한 실험들을 통해, 제안 모델의 효과와 우수성을 확인하였다.