• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.579-582
• /
• 2022

청소년의 성격유형을 분석할 때 소셜미디어 데이터를 활용하여 텍스트 처리로 분석하는 연구는 많이 알려져 있다. 그러나 이미지를 사용하여 성격유형을 분석한 연구는 미비하다. 본 연구는 청소년의 발테그 그림검사로 표현된 이미지를 데이터로 사용하고, CNN을 활용하여 MMTIC의 16가지 청소년의 성격유형을 예측한다. 연구 대상은 중학교 재학생을 대상으로 한다. MMTIC에서 U-band를 제외한 340명의 학생으로 2012년 4월부터 2013년 3월까지 조사하였다. 연구 결과 CNN을 사용하였을 때 21.6% 예측율을 보였으며, CNN Ensemble을 적용하였을 때 23.1%로 2.5%가 증가한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.64-67
• /
• 2021

가짜뉴스는 뉴스 기사 형식을 갖는 날조된 정보를 의미하며, 최근 모바일 인터넷 장치의 보급과 소셜 네트워크 서비스의 대중화로 온라인 확산이 가속화되고 있다. 기존 연구는 가짜뉴스의 탐지를 위해 뉴스의 주제목, 부제목, 리드, 본문 등 뉴스 기사를 이루는 구성요소를 비롯하여 언론사, 기자, 날짜, 확산 경로 등의 메타 데이터를 대상으로 분석하였다. 그러나 뉴스의 제목과 본문 및 메타 데이터 등은 내용 수정이 쉬워, 다량의 데이터를 학습한 모델이라 하더라도 높은 정확도를 장기간 유지하기 어려울 수 있다. 이러한 문제점을 해결하기 위하여 본 논문은 합성곱 신경망을 이용해 문맥 정보를 분석하고 장단기 메모리 기반의 감성분석을 추가로 수행한다. 문맥 정보와 가짜뉴스 유포자가 쉽게 수정할 수 없는 감성 변화 패턴을 활용하여 성능이 개선된 가짜뉴스 탐지 모델을 제안한다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.18 no.2
• /
• pp.157-162
• /
• 2008

This paper presents a learning method of tagging ontology using large tagging data such as a folksonomy, which stands for classification structure informally created by the people. There is no common agreement about the semantics of a tagging, and most social web sites internally use different methods to represent tagging information, obstructing interoperability between sites and the automated processing by software agents. To solve this problem, we need a tagging ontology, defined by analyzing intrinsic attributes of a tagging. Through several machine learning for tagging data, tag groups and similar user groups are extracted, and then used to learn the tagging ontology. A recommender system adopting the tagging ontology is also suggested as an applying field.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.2
• /
• pp.408-414
• /
• 2016

This study will find meaningful independent variables for criterion variables that affect influencing on purchase intention in social network service, on the basis of the results of a meta-analysis. We reviewed a total of 29 studies related purchase intention in social network service published in Korea journals between 2005 and 2015, where a cause and effect relationship is established between variables that are specified in the conceptual model of this study. The result of the meta-analysis might be summarized that the highest effect size (r = .455) is the path from the satisfaction to the purchase intention. The second biggest effect size (r = .398) was found in the path between the word of mouth to the purchase intention. Next, the effect size (r = .386) in the path from the trust to the purchase intention showed very lower. Finally, the result of the meta analysis can be concluded that lower effect size (r = .342) Further, the predictive variables of this study have power of explanation about 22%-12% or more. Based on these findings, several theoretical and practical implications were suggested and discussed.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.909-911
• /
• 2020

소셜 네트워크 서비스 등의 발전으로 인해 개인이 다수에게 의견을 표출하는 통로가 활성화되었다. 게시물에 드러난 감정을 통해 특정 주제에 대한 여론을 도출할 수 있다. 본 논문에서는 BERT를 통한 문장 분석 기술, 그 중에서도 감정 분석을 하는 방법을 분석하고, 이를 일반화된 문장에 적용시키기 위한 데이터 셋 구성에 관하여 연구를 진행하였다.

• The Journal of the Convergence on Culture Technology
• /
• v.8 no.2
• /
• pp.209-215
• /
• 2022

In this study, detailed factors of public sentiment toward the 'post-corona new normal' were examined through social media big data sentiment analysis. Thus, it is to provide basic data to preemptively cope with the post-COVID-19 era. For data collection and analysis, the emotional analysis program of 'Textom', a big data analysis program, was used. The data collection period is one year from October 5, 2020 to October 5, 2021, and the collection channels are set as blogs, cafes, Twitter, and Facebook on Daum and Naver. The original data edited and refined a total of 3,770 collected texts from this channel were used for this study. The conclusion is as follows. First, there is a high level of interest and liking for the 'post-corona new normal'. In other words, it can be seen that optimism such as daily recovery, technological growth, and expectations for a new future took the lead at 77.62%. Second, negative emotions such as sadness and rejection are 22.38% of the total, but the intensity of emotions is 23.91%, which is higher than the ratio, suggesting that these negative emotions are intense. This study has a contribution to the detailed factor analysis of the public's positive and negative emotions through big data analysis on the 'post-corona new normal'.

• Journal of Korean Library and Information Science Society
• /
• v.43 no.4
• /
• pp.347-372
• /
• 2012

This paper explores the current condition of facebook pages in order to provide university libraries with practical ideas in using facebook services. A total of 20 facebook pages of 20 selected Korean and USA university libraries were examined and analyzed in various ways. The research questions for this study are: (1) how well do library operate the facebook page? (2) what are the category and nature of wall posts of library facebook pages. To get the answers to this questions, various checkpoints are developed and applied. The most significant findings are that there are a quite different data of operation and there are similar patterns of posts communication to the Korean and USA university library pages.

• Journal of the Korean Society for information Management
• /
• v.39 no.3
• /
• pp.69-97
• /
• 2022

Parents' interest and care are very important for their children in the process of growing up, but parenting is also very unfamiliar and difficult for themselves. Parents constantly need information related to parenting, and in order to solve this problem, they try to obtain information through informal meetings or social relationships among parents at an individual level, which often brings up conflicts and stresses. Accordingly, the purpose of this study is to explore the characteristics of social networks among parents and to understand the influence of the network characteristics on information-sharing intention and information-sharing satisfaction. To this end, a survey questionnaire was developed based on previous studies and was administered among parents whose children are in a daycare center. Then, the data was analyzed using social network analysis (SNA) method. Findings indicate that the correlation among the social networks, the information -sharing intention, and the information-sharing satisfaction was significant, but the influence of the social networks on the information-sharing intention, and the information-sharing satisfaction was not established. It was also found that the intention to receive help was higher than the desire to help others. Policy suggestions to meet parents' information needs and the need to reconsider the role of local public libraries were presented.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.109-109
• /
• 2022

유효가뭄지수(Effective Drought Index, EDI)는 현재 기후에 대한 과거, 미래 기후에서의 상대적인 가뭄 특성 변화를 평가하기 위해 최근 30년간의 일 강수량 자료를 고정된 기준치로 사용한다. 이에 따라 장기간에 걸친 다양한 기후변화에 대한 인간 사회의 적응을 고려하며 가뭄의 영향을 평가할 때에는 한계점이 있다. 이 연구에서는 EDI의 기능을 확장하기 위해 자가교정 유효가뭄지수(self-calibrating EDI, scEDI)를 제안하고 그 성능을 평가하였다. 기존 EDI와는 다르게 scEDI는 관측시점을 기준으로 30년씩 이동하며 시간에 따라 변화하는 기후값을 기준치로 사용한다. 우리나라 서울관측소에서 1777년부터 2020년까지 누적된 244년간의 일 강수량 자료를 바탕으로, scEDI와 3개의 서로 다른 기후 기준치를 가진 기존 EDI를 계산하여 평가된 가뭄들의 특성을 비교하였다. 그 결과, 기후 기준치에 따라 서로 다른 가뭄 특성들을 보인 기존 EDI와 달리, scEDI는 변화하는 기후를 고려하여 분석기간에 걸쳐 가뭄의 특성을 일관되게 평가할 수 있는 것으로 밝혀졌다. 1807년부터 1907년까지 가뭄과 관련된 조선왕조실록의 기록과 scEDI가 평가한 과거 가뭄 사례들과 비교해 본 결과, scEDI가 탐지한 가뭄 사례들과 실제 조선왕조실록의 가뭄 기록이 비교적 잘 일치하여, scEDI가 과거의 사회적 가뭄을 잘 탐지하는 것으로 나타났다. 또한, 최근의 사회적 가뭄에 대한 scEDI의 탐지 능력을 평가하기 위해 구글과 네이버에서 2016년부터 2018년까지 수집된 가뭄 관련 검색어 소셜 빅데이터를 사용하여서 비교하였다. 그 결과, 과거와 마찬가지로 현재에서도 scEDI가 평가한 가뭄의 변화와 소셜 빅데이터에서 나타난 가뭄에 대한 사회적 반응이 잘 일치하는 것으로 나타났다.

• Information Systems Review
• /
• v.17 no.2
• /
• pp.179-192
• /
• 2015

Collaborative research has been actively done on the basis of academic relationships among various study area. The importance of collaboration has also been increased. Collaborative researchers can reduce time, cost, and research risk to maximize research productivity. This study aims to develop a framework for understanding the behavior of professional groups through network characteristics. To achieve the goal, we collected data of the co-authored network and that of the reviewer network from from 2006 to 2012. Total 230 submitted papers were analyzed on the views of research performance and productivity. Various analytical methods such as centrality analysis, sub-group analysis, correlation, and regression were conducted for assuring the reliability and validity of our research. The results shows that the productivity of the co-authored network was increased and the efficiency of the reviewer network was also identified through several network indexes.