• The Journal of Society for e-Business Studies
• /
• v.26 no.3
• /
• pp.67-80
• /
• 2021

The node embedding technique for learning graph representation plays an important role in obtaining good quality results in graph mining. Until now, representative node embedding techniques have been studied for homogeneous graphs, and thus it is difficult to learn knowledge graphs with unique meanings for each edge. To resolve this problem, the conventional Triple2Vec technique builds an embedding model by learning a triple graph having a node pair and an edge of the knowledge graph as one node. However, the Triple2 Vec embedding model has limitations in improving performance because it calculates the relationship between triple nodes as a simple measure. Therefore, this paper proposes a feature extraction technique based on a graph convolutional neural network to improve the Triple2Vec embedding model. The proposed method extracts the neighborliness vector of the triple graph and learns the relationship between neighboring nodes for each node in the triple graph. We proves that the embedding model applying the proposed method is superior to the existing Triple2Vec model through category classification experiments using DBLP, DBpedia, and IMDB datasets.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.524-526
• /
• 2024

컴퓨터 과학(CS) 분야는 다른 학문 분야에 비해 연구 동향이 빠르게 변하는 특성을 가지고 있다. 그래프 마이닝에서 활발히 사용되는 CS 분야 논문 데이터셋를(e.g., Cora, Citeseer, DBLP)은 오래된 논문을 중심으로 구성되어 있어 이러한 특성을 제대로 반영하지 못하는 한계가 있다. 따라서 본논문에서는 CS 분야의 최신 트렌드를 반영하는 논문 데이터셋을 제안한다. 이를 위해, 우리는 CS 분야 논문을 활발히 공개하는 플랫폼인 arXiv 에서 2007 년부터 2023 년까지 해당 플랫폼에서 공개된 논문들을 수집하고, 이를 기반으로 공저자 그래프 및 인용 그래프로 구축한다. 해당 데이터셋을 대상으로 폭넓은 분석을 통해, 우리가 구축한 데이터셋이 실세계 그래프 네트워크 특성을 잘 반영하고 있음을 보인다. 또한, 향휴에 해당 데이터셋을 사용하려는 연구자들을 위해, 해당 데이터셋에서의 기존 그래프 기반 응용들의 노드 분류 성능을 제시한다.

• Journal of KIISE:Databases
• /
• v.33 no.7
• /
• pp.754-765
• /
• 2006

In XML keyword search, a search result is defined as a set of the smallest elements (i.e., least common ancestors) containing all query keywords and a granularity of indexing is an XML element instead of a document. Under the conventional index structure, all least common ancestors produced by the combination of the elements, each of which contains a query keyword, are considered as a search result. In this paper, to avoid unnecessary operations of producing the least common ancestors and reduce query process time, we describe a way to construct a partitioned index composed of several partitions and produce a search result by merging those partitions if necessary. When a search result is restricted to be composed of the least common ancestors whose depths are higher than a given minimum depth, under the proposed partitioned index structure, search systems can reduce the query process time by considering only combinations of the elements belonging to the same partition. Even though the minimum depth is not given or unknown, search systems can obtain a search result with the partitioned index, which requires the same query process time to obtain the search result with non-partitioned index. Our experiment was conducted with the XML documents provided by the DBLP site and INEX2003, and the partitioned index could reduce a substantial amount of query processing time when the minimum depth is given.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.12
• /
• pp.5888-5903
• /
• 2019

The Relational ranking method applies authority-based ranking in relational dataset that can be modeled as graphs considering also their tuples' values. Authority directions from tuples that contain the given keywords and transfer to their corresponding neighboring nodes in accordance with their values and semantic connections. From our previous work, ObjectRank extends to ValueRank that also takes into account the value of tuples in authority transfer flows. In a maked difference from ObjectRank, which only considers authority flows through relationships, it is only valid in the bibliographic databases e.g. DBLP dataset, ValueRank facilitates the estimation of importance for any databases, e.g. trading databases, etc. A relational keyword search paradigm Object Summary (denote as OS) is proposed recently, given a set of keywords, a group of Object Summaries as its query result. An OS is a multilevel-tree data structure, in which node (namely the tuple with keywords) is OS's root node, and the surrounding nodes are the summary of all data on the graph. But, some of these trees have a very large in total number of tuples, size-l OSs are the OS snippets, have also been investigated using ValueRank.We evaluated the real bibliographical dataset and Microsoft business databases to verify of our proposed approach.

• Journal of Information Science Theory and Practice
• /
• v.2 no.2
• /
• pp.6-15
• /
• 2014

Initial based disambiguation of author names is a common data pre-processing step in bibliometrics. It is widely accepted that this procedure can introduce errors into network data and any subsequent analytical results. What is not sufficiently understood is the precise impact of this step on the data and findings. We present an empirical answer to this question by comparing the impact of two commonly used initial based disambiguation methods against a reasonable proxy for ground truth data. We use DBLP, a database covering major journals and conferences in computer science and information science, as a source. We find that initial based disambiguation induces strong distortions in network metrics on the graph and node level: Authors become embedded in ties for which there is no empirical support, thus increasing their sphere of influence and diversity of involvement. Consequently, networks generated with initial-based disambiguation are more coherent and interconnected than the actual underlying networks, and individual authors appear to be more productive and more strongly embedded than they actually are.

• Journal of the Korean Society for information Management
• /
• v.35 no.1
• /
• pp.157-182
• /
• 2018

As data management and processing techniques have been developed rapidly in the era of big data, nowadays a lot of business companies and researchers have been interested in long tail data which were ignored in the past. This study proposes methods for generating and controlling a network of technical terms based on text mining technique to enhance data utilization in the distribution of long tail theory. Especially, an edit distance technique of text mining has given us efficient methods to automatically create an interlinking network of technical terms in the scholarly field. We have also used linked open data system to gather experimental data to improve data utilization and proposed effective methods to use data of LOD systems and algorithm to recognize patterns of terms. Finally, the performance evaluation test of the network of technical terms has shown that the proposed methods were useful to enhance the rate of data utilization.

• Proceedings of the Korea Contents Association Conference
• /
• 2008.05a
• /
• pp.823-826
• /
• 2008

This paper shows linkage management for external services. There are many services for specific entities such as DBLP and OntoWorld. OntoFrame, as a Semantic Web-based research information service portal, aims at one-stop service in ways that it connects external services with hyperlinks. For managing the linkage, linkage rules are manually edited by human administrators and automatically verified and tested by linkage management system. It consists of linkage rule management, linkage rule verification, linkage test, and dynamic link generation. Linkage rule management creates and edits linkage rules to connect external services on the Web. After finished rule editing and verification step, linkage management invokes linkage test with entity list. Only valid links are visible to enable users to click on our system, and thus it increases user's reliability on the system.