• 국제학술발표논문집
• /
• The 9th International Conference on Construction Engineering and Project Management
• /
• pp.752-759
• /
• 2022

The instantiation of spaces as a discrete entity allows users to utilize BIM models in a wide range of analyses. However, in practice, their utility has been limited as spaces are erroneously entered due to human error and often omitted entirely. Recent studies attempted to automate space allocation using artificial intelligence approaches. However, there has been limited success as most studies focused solely on the use of geometric features to distinguish spaces. In this study, in addition to geometric features, semantic relations between spaces and elements were modeled and used to improve space classification in BIM models. Graph Convolutional Networks (GCN), a deep learning algorithm specifically tailored for learning in graphs, was deployed to classify spaces via a similarity graph that represents the relationships between spaces and their surrounding elements. Results confirmed that accuracy (ACC) was +0.08 higher than the baseline model in which only geometric information was used. Most notably, GCN was able to correctly distinguish spaces with no apparent difference in geometry by discriminating the specific elements that were provided by the similarity graph.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권11호
• /
• pp.2903-2923
• /
• 2023

Knowledge recommendation is a type of recommendation system that recommends knowledge content to users in order to satisfy their needs. Although using graph neural networks to extract data features is an effective method for solving the recommendation problem, there is information loss when modeling real-world problems because an edge in a graph structure can only be associated with two nodes. Because one super-edge in the hypergraph structure can be connected with several nodes and the effectiveness of knowledge graph for knowledge expression, a dual-channel hypergraph convolutional neural network model (DCHC) based on hypergraph structure and knowledge graph is proposed. The model divides user data and knowledge data into user subhypergraph and knowledge subhypergraph, respectively, and extracts user data features by dual-channel hypergraph convolution and knowledge data features by combining with knowledge graph technology, and finally generates recommendation results based on the obtained user embedding and knowledge embedding. The performance of DCHC model is higher than the comparative model under AUC and F1 evaluation indicators, comparative experiments with the baseline also demonstrate the validity of DCHC model.

• 한국ITS학회 논문지
• /
• 제20권1호
• /
• pp.70-85
• /
• 2021

최근 활발히 연구되는 딥러닝 방법론은 인공지능의 성능을 급속도로 향상시켰고, 이에 따라 다양한 산업 분야에서 딥러닝을 활용한 시스템이 제시되고 있다. 교통 시스템에서는 GNN을 활용한 공간-시간 그래프 모델링이 교통 속도 예측에 효과적인 것으로 밝혀졌지만, 이는 메모리 병목 현상을 유발하기 때문에 모델이 비효율적으로 학습된다는 단점이 있다. 따라서 본 연구에서는 그래프 분할 방법을 통해 도로 네트워크를 분할하여 메모리 병목 현상을 완화함과 동시에 우수한 성능을 달성하고자 한다. 제안 방법론을 검증하기 위해 인천시 UTIC 데이터 분석 결과를 바탕으로 Jensen-Shannon divergence를 사용하여 도로 속도 분포의 유사도를 측정하였다. 그리고 측정된 유사도를 바탕으로 스펙트럴 클러스터링을 수행하여 도로 네트워크를 군집화하였다. 성능 측정 결과, 도로 네트워크가 7개의 네트워크로 분할되었을 때 MAE 기준 5.52km/h의 오차로 비교 모델 대비 가장 우수한 정확도를 보임과 동시에 메모리 병목 현상 또한 완화되는 것을 확인할 수 있었다.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제11권4호
• /
• pp.19-38
• /
• 2007

Graph theory is becoming increasingly significant as it is applied of mathematics, science and technology. It is being actively used in fields as varied as biochemistry(genomics), electrical engineering(communication networks and coding theory), computer science(algorithms and computation) and operations research(scheduling). The powerful results in other areas of pure mathematics. Rhis paper, besides giving a general outlook of these facts, includes new graph theoretical proofs of Fermat's Little Theorem and the Nielson-Schreier Theorem. New applications to DNA sequencing (the SNP assembly problem) and computer network security (worm propagation) using minimum vertex covers in graphs are discussed. We also show how to apply edge coloring and matching in graphs for scheduling (the timetabling problem) and vertex coloring in graphs for map coloring and the assignment of frequencies in GSM mobile phone networks. Finally, we revisit the classical problem of finding re-entrant knight's tours on a chessboard using Hamiltonian circuits in graphs.

• International Journal of Contents
• /
• 제17권3호
• /
• pp.1-14
• /
• 2021

Since the outbreak of the coronavirus disease 2019 (COVID-19) pandemic, a lot of efforts have been made in the field of data science to help combat against this disease. Among them, forecasting the number of cases of infection is a crucial problem to predict the development of the pandemic. Many deep learning-based models can be applied to solve this type of time series problem. In this research, we would like to take a step forward to incorporate spatial data (geography) with time series data to forecast the cases of region-level infection simultaneously. Specifically, we model a single spatio-temporal graph, in which nodes represent the geographic regions, spatial edges represent the distance between each pair of regions, and temporal edges indicate the node features through time. We evaluate this approach in COVID-19 in a Korean dataset, and we show a decrease of approximately 10% in both RMSE and MAE, and a significant boost to the training speed compared to the baseline models. Moreover, the training efficiency allows this approach to be extended for a large-scale spatio-temporal dataset.

• 한국멀티미디어학회논문지
• /
• 제24권3호
• /
• pp.423-430
• /
• 2021

Aiming Graph convolutional networks (GCNs) have achieved outstanding performances on skeleton-based action recognition. However, several problems remain in existing GCN-based methods, and the problem of low recognition rate caused by single input data information has not been effectively solved. In this article, we propose a Dual-stream fusion method that combines video data and skeleton data. The two networks respectively identify skeleton data and video data and fuse the probabilities of the two outputs to achieve the effect of information fusion. Experiments on two large dataset, Kinetics and NTU-RGBC+D Human Action Dataset, illustrate that our proposed method achieves state-of-the-art. Compared with the traditional method, the recognition accuracy is improved better.

• International journal of advanced smart convergence
• /
• 제12권4호
• /
• pp.88-97
• /
• 2023

Traffic flow prediction is of great significance in urban planning and traffic management. As the complexity of urban traffic increases, existing prediction methods still face challenges, especially for the fusion of spatiotemporal information and the capture of long-term dependencies. This study aims to use the fusion model of graph neural network to solve the spatio-temporal information fusion problem in traffic flow prediction. We propose a new deep learning model Spatio-Temporal Information Fusion using Graph Neural Networks (STFGNN). We use GCN module, TCN module and LSTM module alternately to carry out spatiotemporal information fusion. GCN and multi-core TCN capture the temporal and spatial dependencies of traffic flow respectively, and LSTM connects multiple fusion modules to carry out spatiotemporal information fusion. In the experimental evaluation of real traffic flow data, STFGNN showed better performance than other models.

• 정보과학회 논문지
• /
• 제43권3호
• /
• pp.380-388
• /
• 2016

그래프 클러스터링은 서로 유사한 특성을 갖는 정점들을 동일한 클러스터로 묶는 기법으로 그래프 데이터를 분석하고 그 특성을 파악하는데 폭넓게 사용된다. 최근 소셜 네트워크 서비스와 월드 와이드 웹, 텔레폰 네트워크 등의 다양한 응용분야에서 크기가 큰 대용량 그래프 데이터가 생성되고 있다. 이에 따라서 대용량 그래프 데이터를 효율적으로 처리하는 클러스터링 기법의 중요성이 증가하고 있다. 본 논문에서는 대용량 그래프 데이터의 클러스터들을 효율적으로 생성하는 클러스터링 알고리즘을 제안한다. 우리의 제안 기법은 그래프 내의 클러스터들 간의 유사도를 Min-Hash를 이용하여 효과적으로 추정하고 계산된 유사도에 따라서 클러스터들을 생성한다. 실세계 데이터를 이용한 실험에서 우리는 본 논문에서 제안하는 기법과 기존 그래프 클러스터링 기법들과 비교하여 제안기법의 효율성을 보였다.

• 한국생물정보학회:학술대회논문집
• /
• 한국생물정보시스템생물학회 2005년도 BIOINFO 2005
• /
• pp.267-271
• /
• 2005

The goal of data mining is to extract new and useful knowledge from large scale datasets. As the amount of available data grows explosively, it became vitally important to develop faster data mining algorithms for various types of data. Recently, an interest in developing data mining algorithms that operate on graphs has been increased. Especially, mining frequent patterns from structured data such as graphs has been concerned by many research groups. A graph is a highly adaptable representation scheme that used in many domains including chemistry, bioinformatics and physics. For example, the chemical structure of a given substance can be modelled by an undirected labelled graph in which each node corresponds to an atom and each edge corresponds to a chemical bond between atoms. Internet can also be modelled as a directed graph in which each node corresponds to an web site and each edge corresponds to a hypertext link between web sites. Notably in bioinformatics area, various kinds of newly discovered data such as gene regulation networks or protein interaction networks could be modelled as graphs. There have been a number of attempts to find useful knowledge from these graph structured data. One of the most powerful analysis tool for graph structured data is frequent subgraph analysis. Recurring patterns in graph data can provide incomparable insights into that graph data. However, to find recurring subgraphs is extremely expensive in computational side. At the core of the problem, there are two computationally challenging problems. 1) Subgraph isomorphism and 2) Enumeration of subgraphs. Problems related to the former are subgraph isomorphism problem (Is graph A contains graph B?) and graph isomorphism problem(Are two graphs A and B the same or not?). Even these simplified versions of the subgraph mining problem are known to be NP-complete or Polymorphism-complete and no polynomial time algorithm has been existed so far. The later is also a difficult problem. We should generate all of 2$^n$ subgraphs if there is no constraint where n is the number of vertices of the input graph. In order to find frequent subgraphs from larger graph database, it is essential to give appropriate constraint to the subgraphs to find. Most of the current approaches are focus on the frequencies of a subgraph: the higher the frequency of a graph is, the more attentions should be given to that graph. Recently, several algorithms which use level by level approaches to find frequent subgraphs have been developed. Some of the recently emerging applications suggest that other constraints such as connectivity also could be useful in mining subgraphs : more strongly connected parts of a graph are more informative. If we restrict the set of subgraphs to mine to more strongly connected parts, its computational complexity could be decreased significantly. In this paper, we present an efficient algorithm to mine frequent subgraphs that are more strongly connected. Experimental study shows that the algorithm is scaling to larger graphs which have more than ten thousand vertices.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권1호
• /
• pp.421-433
• /
• 2015

Recently, there have been fast-growing social network services based on the Internet environment and web technology development, the prevalence of smartphones, etc. Social networks also allow the users to convey the information and news so that they have a great influence on the public opinion formed by social interaction among users as well as the spread of information. On the other hand, these social networks also serve as perfect environments for rampant malware. Malware is rapidly being spread because relationships are formed on trust among the users. In this paper, an effective patch strategy is proposed to deal with malicious worms based on social networks. A graph is formed to analyze the structure of a social network, and subgroups are formed in the graph for the distributed patch strategy. The weighted directions and activities between the nodes are taken into account to select reliable key nodes from the generated subgroups, and the Incremental PageRanking algorithm reflecting dynamic social network features (addition/deletion of users and links) is used for deriving the high influential key nodes. With the patch based on the derived key nodes, the proposed method can prevent worms from spreading over social networks.