UML 2.0 행위 다이어그램을 확장한 비동기 상호작용 웹 애플리케이션의 항해 모델

변창현(ChangHyun Byun),서정호(JungHo Seo),류성열(SungYul Rhew) 한국정보과학회 2007 한국정보과학회 학술발표논문집 Vol.34 No.2B

Web2.0 의 비동기식 웹 페이지 전환기술이 주목을 받으면서 기존의 웹사이트들이 동적인 웹 사이트로 바뀌어 가고 있다. 그러나 웹 애플리케이션 개발에 적용 가능한 설계모델인 UML2.0은 정적인 웹 페이지 설계 모델로 국한되어 있다. 따라서 본 논문에서는 UWE의 항해모델에서 뷰 관점 항해모델과, 데이터 전송 항해 모델을 State Machine, Communication Diagram을 이용하여 비동기식 웹 페이지 항해 모델 개선과 기존의 동기식 웹 페이지 모델과의 호환성 방안을 제안한다.

시공간적 인공신경망을 활용한 서울시 공기질 변화 예측

변종윤(Byun, JongYun),전창현(Jun, Changhyun) 대한토목학회 2021 대한토목학회 학술대회 Vol.2021 No.10

Bioinspired, Calcium-Free Alginate Hydrogels with Tunable Physical and Mechanical Properties and Improved Biocompatibility

Lee, Changhyun,Shin, Jisoo,Lee, Jung Seung,Byun, Eunkyoung,Ryu, Ji Hyun,Um, Soong Ho,Kim, Dong-Ik,Lee, Haeshin,Cho, Seung-Woo American Chemical Society 2013 Biomacromolecules Vol.14 No.6

<P>Alginate hydrogels are for various biomedical applications including tissue engineering, cell therapy, and drug delivery. However, it is not easy to control swelling or viscoelastic and biophysical properties of alginate hydrogels prepared by conventional cross-linking methods (ionic interaction using divalent cations). In this study, we describe a bioinspired approach for preparing divalent ion-free alginate hydrogels that exhibit tunable physical and mechanical properties and improved biocompatibility due to the absence of cations in the gel matrices. We conjugated dopamine, a minimalized adhesive motif found in the holdfast pads of mussels, to alginate backbones (alginate-catechol) and the tethered catechols underwent oxidative cross-linking. This resulted in divalent cation-free alginate hydrogels. The swelling ratios and moduli of the alginate-catechol hydrogels are readily tunable, which is difficult to achieve in ionic bond-based alginate hydrogels. Furthermore, alginate-catechol hydrogels enhanced the survival of various human primary cells including stem cells in the three-dimensional gel matrix, indicating that intrinsic cytotoxicity caused by divalent cations becomes negligible when employing catechol oxidation for alginate cross-linking. The inflammatory response in vivo was also significantly attenuated compared to conventional alginate hydrogels with calcium cross-linking. This biomimetic approach for the preparation of alginate hydrogels may provide a novel platform technology to develop tunable, functional, biocompatible, three-dimensional scaffolds for tissue engineering and cell therapy.</P><P><B>Graphic Abstract</B> <IMG SRC='http://pubs.acs.org/appl/literatum/publisher/achs/journals/content/bomaf6/2013/bomaf6.2013.14.issue-6/bm400352d/production/images/medium/bm-2013-00352d_0009.gif'></P>

The Influence of Social Media: Twitter Usage Pattern during the 2014 Super Bowl Game

Hyeonjeong Shin,Changhyun Byun,Hyeoncheol Lee 보안공학연구지원센터 2015 International Journal of Multimedia and Ubiquitous Vol.10 No.3

In recent years, with the increasing popularity of diverse online social network sites, such as Facebook, Twitter, Blogger, YouTube, LinkedIn, and MySpace, a massive amount of data has become available. Applying data mining techniques to social media can yield interesting perspectives to understanding individual and human behavior, detecting hot issues and topics, or discovering a group and community. This study aims to address the question of how people use Twitter and to assess the power of Twitter in terms of creating consumer interest in commercials based on an analysis of Twitter messages about various Super Bowl ads. Instead of relying on the audience’s responses, this study adopts a hybrid approach, combining a data-mining approach to collect, filter and analyze a massive amount of tweets with manual content analysis.

A Resource-Efficient System Architecture for Processing Various Sensor Data in Smart home Environment

Changmin Lee,Changhyun Byun,Hyeonjeong Shin 보안공학연구지원센터 2016 International Journal of Smart Home Vol.10 No.11

The Internet of Things (IoT) technology is gaining vast popularity and it is expected to become ubiquitous in the near future. It envisions the idea of a fully connected network of smart objects, enabling cooperative and intelligent distributed functionalities. Especially, smart homes provide innovative, automated and interactive services for residential customers through distributed and collaborative operations. In this paper the challenge of processing large amounts of sensor data at network gateways with limited memory and computing power is described. A resource-efficient system to process various heterogeneous data is introduced.

격자 기반 공간보간 기법을 활용한 국내 PM10 자료의 시공간적 변동성 분석

이진욱(Lee, Jinwook),변종윤(Byun, Jongyun),황승현(Hwang, Seunghyun),전창현(Jun, Changhyun),백종진(Baik, Jongjin) 한국방재학회 2022 한국방재학회논문집 Vol.22 No.1

본 연구에서는 2001년부터 2019년까지 관측된 국내 PM10 자료에 대한 시공간적 변동성 분석을 수행하였다. 이를 위해 연도별로 최소 175개에서 최대 484개의 지점 자료를 역거리가중법을 이용하여 공간보간하였다. 웨이블릿 분석을 통해 일평균 자료에 대한 주기적 변동성에 대해 검토한 결과, PM10은 약 1년 정도의 기간을 두고 규칙적으로 반복하는 것을 확인할 수 있었다. 월평균 및 연평균 자료에 대해 Mann-Kendall 검정을 적용한 결과, 1년에 약 1 μg/m3 정도 감소하는 추세를 보임을 확인하였다. 연평균 자료를 이용하여 격자 단위의 공간적 변화를 살펴본 결과, 북쪽으로 갈수록 상대적으로 PM10이 높게 나타났으며 시간이 지남에 따라 절대적인 관측값 및 편차가 상당히 감소된 것을 확인할 수 있었다. 전체 기간에 대한 연평균 자료들의 평균은 수도권 부근이 상대적으로 높은 것으로 확인되었으며, 표준편차에서도 유사한 경향을 확인할 수 있었다. This study analyzed spatiotemporal variability in domestic PM10 data from 2001 to 2019. From annual numbers of stations between 175 and 484, the point data at each station were spatially interpolated using the inverse distance weighted method. A periodic variability in daily mean data was examined through wavelet analysis, which showed that there was a clear annual pattern with the periodic change following a regular cycle. The Mann-Kendall Test for monthly and annual mean data showed a decreasing trend in about 1 μg/m3 per year. The spatial change in the grid data for annual mean data represented that it was relatively higher in the northern regions than that in the southern regions and its mean and deviation decreased significantly over time. For the entire period of observation data, it was found that annual mean and standard deviation of PM10 concentrations were relatively high in the region near the metropolitan area.

Dynamic Seed Selection for Twitter Data Collection

Hyoenchoel Lee(이현철),Changhyun Byun(변창현),Yanggon Kim(김양곤),Sang Ho Lee(이상호) 한국정보과학회 2014 정보과학회논문지 : 데이타베이스 Vol.41 No.4

트위터와 같은 소셜 네트워크 분석은 인간의 행동을 이해하거나, 화제가 되는 주제를 탐지하거나, 영향력 있는 사람을 식별하거나, 커뮤니티나 그룹을 발견하는데 흥미로운 시각을 제공할 수 있다. 하지만 소셜 네트워크가 가지는 특성(즉 데이터가 방대하고, 정교하지 않으며 또한 동적인 특성) 으로 인하여 소셜 네트워크에서 주제와 연관이 있는 데이터를 수집하는 것은 어려운 일이다. 본 논문은 주어진 주제와 관련 있는 트윗을 효과적으로 수집하기 위하여 시드 노드를 동적으로 선택하는 알고리즘을 제안한다. 본 알고리즘은 사용자의 영향력을 측정하기 위하여 사용자 속성을 활용하며, 수집 프로세스 중에 시드 노드를 동적으로 할당한다. 우리는 제안한 알고리즘을 실제 트윗 데이터에 적용하였으며, 만족할 만한 성능결과를 얻었다. Analysis of social media such as Twitter can yield interesting perspectives to understanding human behavior, detecting hot issues, identifying influential people, or discovering a group and community. However, it is difficult to gather the data relevant to specific topics due to the main characteristics of social media data; data is large, noisy, and dynamic. This paper proposes a new algorithm that dynamically selects the seed nodes to efficiently collect tweets relevant to topics. The algorithm utilizes attributes of users to evaluate the user influence, and dynamically selects the seed nodes during the collection process. We evaluate the proposed algorithm with real tweet data, and get satisfactory performance results.

Two-Dimensional Materials Inserted at the Metal/Semiconductor Interface: Attractive Candidates for Semiconductor Device Contacts

Lee, Min-Hyun,Cho, Yeonchoo,Byun, Kyung-Eun,Shin, Keun Wook,Nam, Seong-Geol,Kim, Changhyun,Kim, Haeryong,Han, Sang-A,Kim, Sang-Woo,Shin, Hyeon-Jin,Park, Seongjun American Chemical Society 2018 NANO LETTERS Vol.18 No.8

<P>Metal-semiconductor junctions are indispensable in semiconductor devices, but they have recently become a major limiting factor precluding device performance improvement. Here, we report the modification of a metal/n-type Si Schottky contact barrier by the introduction of two-dimensional (2D) materials of either graphene or hexagonal boron nitride (h-BN) at the interface. We realized the lowest specific contact resistivities (ρ<SUB>c</SUB>) of 3.30 nΩ cm<SUP>2</SUP> (lightly doped n-type Si, ∼ 10<SUP>15</SUP>/cm<SUP>3</SUP>) and 1.47 nΩ cm<SUP>2</SUP> (heavily doped n-type Si, ∼ 10<SUP>21</SUP>/cm<SUP>3</SUP>) via 2D material insertion are approaching the theoretical limit of 1.3 nΩ cm<SUP>2</SUP>. We demonstrated the role of the 2D materials at the interface in achieving a low ρ<SUB>c</SUB> value by the following mechanisms: (a) 2D materials effectively form dipoles at the metal-2D material (M/2D) interface, thereby reducing the metal work function and changing the pinning point, and (b) the fully metalized M/2D system shifts the pinning point toward the Si conduction band, thus decreasing the Schottky barrier. As a result, the fully metalized M/2D system using atomically thin and well-defined 2D materials shows a significantly reduced ρ<SUB>c</SUB>. The proposed 2D material insertion technique can be used to obtain extremely low contact resistivities in metal/n-type Si systems and will help to achieve major performance improvements in semiconductor technologies.</P> [FIG OMISSION]</BR>

Study on Efficient Topic Detection Method through Topic Modeling Approaches

Hyunjeong Shin(신현정),Changmin Lee(이창민),Changhyun Byun(변창현) 융복합지식학회 2018 융복합지식학회논문지 Vol.6 No.2

스택 오버플로우, 애스크 우분투, 야수 앤서, 쿼오라 같은 유저가 질문을 올리고 질문에 답변을 다는 온라인 커뮤니티 사이트들을 지식 공유 플랫폼이라 한다. 이때 각 질문에 사용자들이 태그를 추가할 수 있다. 태깅은 전반적인 온라인 커뮤니티 내에서는 매우 인기가 많은 시스템으로 태킹을 함으로서 사용자들이 자신들이 올리는 질문을 분류할 수 있어, 플랫폼에서는 비용을 들여 전문가를 고용하여 질문들을 별도로 분류를 할 필요가 없다는 장점이 있다. 그러나 태깅은 그 자체의 장점 뿐만 아니라 문제점도 가지도 있는데 태그 폭발과 태그 동의어가 대표적이다. 이 논문에서는 지식 공유 플랫폼에서 토픽 모델링 기법을 사용한 태그 추천 시스템을 제안한다. 토픽모델링은 문사 유사도와 차원 감소의 다양한 장점들을 포함하고 있으며, 논문에서 제안하는 태그 추천 시스템에서는 문사 유사도와 기록에 기반한 태그 발현 횟수를 계산하여 사용한다. Online community sites such as Stack Overflow, Ask Ubuntu, Yahoo Answer, and Quora are knowledge sharing platforms that allow users to post questions, answer the questions asked by other users, and add tags to their questions. Tagging is a famous system across online communities because tagging make users to classify their questions. It is much cheaper than hiring expert to categorize questions. However, tagging comes with the problem of the tag explosion and the tag synonym. In this paper, we propose a tag recommendation system using topic modeling approaches on knowledge sharing platforms to solve these problems. Topic models have advantages of dimensionality reduction and document similarity. Our tag recommendation system considers the document similarity and the historical tag occurrence to calculate tag scores.

Numerical Analysis on the Cooling Performance of Triply Periodic Minimal Surfaces Cooling Channels under Various Operating Conditions

Hyejun You(유혜준),Sewon Lee(이세원),Seonwoong Byun(변선웅),Changhyun Baek(백창현),Yongchan Kim(김용찬) 대한기계학회 2023 대한기계학회 춘추학술대회 Vol.2023 No.4