• The Journal of the Korea Contents Association
• /
• v.20 no.10
• /
• pp.119-133
• /
• 2020

Recently, as famous YouTubers produce and broadcast videos that receive sponsorship and advertising such as indirect advertising (PPL), a so-called 'back advertising' controversy continues, and not only famous YouTubers but also entertainers are caught up in the issue. It is causing confusion among the public in Korea. This study attempts to find out the public's reaction before and after the controversy of 'back advertising' by YouTubers through comment analysis. Specifically, among text analysis using R programs, we intend to analyze the issue through various methods such as word cloud, qgraph analysis, LDA, and word2vec analysis, a deep learning technique. The target of the analysis was to analyze the channels of three YouTubers who belonged to the controversy of the 'back advertising' YouTuber and uploaded the 'Apology video'. The 5 most recent videos of Muk-bang YouTuber Moon Bok-hee, who has a similar content disposition to SussTV's Han Hye-yeon stylist, which was controversial, and Yang Pang, a YouTuber who showed various contents (August 09, 2020) Criterion and her first 5 videos uploaded were reviewed. As a result of the study, most of the comments that showed positive reactions before the controversy, but after the controversy, it was found that negative reactions accounted for most of the comments. Therefore, this study examines the degree of change of the public about influencers through comments after the controversy over 'back advertising' through various analysis using R program. This research also devises various measures to prevent the occurrence of back advertising of influencers in the future.

• Journal of the Korean Society for information Management
• /
• v.39 no.1
• /
• pp.257-280
• /
• 2022

In this study, we analyzed comments on news articles of representative companies of the three industries (i.e., semiconductor, secondary battery, and bio industries) that had been listed as national strategic technology projects of South Korea to identify public opinions towards them. In addition, we analyzed the relationship between changes in public opinion and stock price. 'Samsung Electronics' and 'SK Hynix' in the semiconductor industry, 'Samsung SDI' and 'LG Chem' in the secondary battery industry, and 'Samsung Biologics' and 'Celltrion' in the bio-industry were selected as the representative companies and 47,452 comments of news articles about the companies that had been published from January 1, 2020, to December 31, 2020, were collected from Naver News. The comments were grouped into positive, neutral, and negative emotions, and the dynamic topics of comments over time in each group were analyzed to identify the trends of public opinion in each industry. As a result, in the case of the semiconductor industry, investment, COVID-19 related issues, trust in large companies such as Samsung Electronics, and mention of the damage caused by changes in government policy were the topics. In the case of secondary battery industries, references to investment, battery, and corporate issues were the topics. In the case of bio-industries, references to investment, COVID-19 related issues, and corporate issues were the topics. Next, to understand whether the sentiment of the comments is related to the actual stock price, for each company, the changes in the stock price and the sentiment values of the comments were compared and analyzed using visual analytics. As a result, we found a clear relationship between the changes in the sentiment value of public opinion and the stock price through the similar patterns shown in the change graphs. This study analyzed comments on news articles that are highly related to stock price, identified changes in public opinion trends in the COVID-19 era, and provided objective feedback to government agencies' policymaking.

• The Journal of the Korea Contents Association
• /
• v.22 no.1
• /
• pp.74-85
• /
• 2022

Along with the change in the values of marriage and the prevalence of non-marriage in Korean society, a new form of family composition called unmarried birth or non-maritalism childbirth has appeared, and social discussion in taking place in connection with the problem of a decrease in the birthrate. Using sentiment analysis and social network analysis, this research explored how the people's sentiment and perception has changed toward 'nonmarital birth.' The data used is comments on news articles from the period of November 2020 to August 2021. As a result of the study, there were a lot of positive comments during the social issue period by marriage, whereas there were many negative comments from the policy agenda to the policy making period. As a result of co-occurrence network analysis, the topic of family norm, policy, and personal aspect appeared. This study is significant in that it revealed that negative perceptions prevailed during the policy-making process after the issue of unmarried births after the issue of unmarried births, and it became a cornerstone of social discussion on unmarried births

• 한국어정보학회:학술대회논문집
• /
• 2016.10a
• /
• pp.291-295
• /
• 2016

본 논문에서는 한국어 기본 유니트 단위로 WPM을 활용한 구글 플레이 스토어 앱의 댓글 감정분석을 수행하였다. 먼저 자동 띄어쓰기 시스템을 적용한 후, 어절단위, 형태소 분석기, WPM을 각각 적용하여 모델을 생성하고, 로지스틱 회귀(Logistic Regression), 소프트맥스 회귀(Softmax Regression), 서포트 벡터머신(Support Vector Machine, SVM)등의 알고리즘을 이용하여 댓글 감정(긍정과 부정)을 비교 분석하였다. 그 결과 어절단위, 형태소 분석기보다 WPM이 최대 25%의 향상된 결과를 얻었다. 또한 분류 과정에서 로지스틱회귀, 소프트맥스 회귀보다는 SVM 성능이 우수했으며, SVM의 기본 파라미터({'kernel':('linear'), 'c':[4]})보다 최적의 파라미터를 적용({'kernel': ('linear','rbf', 'sigmoid', 'poly'), 'C':[0.01, 0.1, 1.4.5]} 하였을 때, 최대 91%의 성능이 나타났다.

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.291-295
• /
• 2016

본 논문에서는 한국어 기본 유니트 단위로 WPM을 활용한 구글 플레이 스토어 앱의 댓글 감정분석을 수행하였다. 먼저 자동 띄어쓰기 시스템을 적용한 후, 어절단위, 형태소 분석기, WPM을 각각 적용하여 모델을 생성하고, 로지스틱 회귀(Logistic Regression), 소프트맥스 회귀(Softmax Regression), 서포트 벡터머신(Support Vector Machine, SVM)등의 알고리즘을 이용하여 댓글 감정(긍정과 부정)을 비교 분석하였다. 그 결과 어절단위, 형태소 분석기보다 WPM이 최대 25%의 향상된 결과를 얻었다. 또한 분류 과정에서 로지스틱회귀, 소프트맥스 회귀보다는 SVM 성능이 우수했으며, SVM의 기본 파라미터({'kernel':('linear'), 'c':[4]})보다 최적의 파라미터를 적용({'kernel': ('linear','rbf', 'sigmoid', 'poly'), 'C':[0.01, 0.1, 1.4.5]} 하였을 때, 최대 91%의 성능이 나타났다.

• Annual Conference on Human and Language Technology
• /
• 2009.10a
• /
• pp.145-150
• /
• 2009

이 논문은 한글 뉴스 기사의 댓글에 대한 감정 분류 방법을 제안한다. 제안된 방법은 기계학습을 이용하는데 본 논문에서는 자질의 가중치를 재조정하는 좀 색다른 방법을 제안한다. 일반적으로 댓글은 독자들이 특정 기사에 대해서 어떠한 감정을 가지고 있는지를 파악하는 중요한 단서가 된다. 그런데 독자들의 감정은 가사에 어떤 분야에 속하느냐에 영향을 받는다. 예를 들면 정치 기사는 부정적인 댓글은 많이 포함하고 있으며 인물 기사는 긍정적인 기사를 많이 포함한다. 이 논문은 이와 같은 댓글의 속성을 이용해서 기사의 원문과 기사의 분야 정보를 이용하여 가중치를 조정한다. 제안된 시스템의 성능을 평가하기 위해 신문 기사와 댓글을 수집하여 감정 말뭉치를 구축하였으며 감정자질을 추출하기 위해 감정 사전을 구축하였다. 제안된 시스템의 $F_1$ 척도는 92.2%였으며 원문의 감정 단어와 분야 정보가 댓글의 감정을 분류하는데 중요한 자질임을 알 수 있었다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2020.07a
• /
• pp.219-222
• /
• 2020

웹툰 시장의 성장에 따라, 웹툰을 주제로 다양한 연구가 진행되고 있다. 그러나 웹툰의 댓글을 분석한 연구는 특정 웹툰에 한정된 연구가 많아 일반적인 웹툰 독자의 댓글 특징을 보기에는 한계가 있다. 또한 웹툰의 흥행과 관련하여, 흥행에 성공한 웹툰과 그렇지 못한 웹툰의 독자 반응을 파악하는 연구는 부족한 실정이다. 따라서 본 연구는 웹툰의 흥행에 주목하여, 흥행의 지표를 웹툰 플랫폼 정식연재로 판단하고, 정식연재가 된 웹툰과 되지 못한 웹툰의 댓글을 비교 분석하였다. 분석 결과, 정식연재가 된 웹툰은 긍정적인 감상평 외에 2차적 저작물을 언급하고 등장인물의 이름 언급이 높았으나 정식연재가 되지 못한 웹툰은 부정적인 감상평 외에, 웹툰 요소에 대한 부정적 언급과 웹툰 장르와 다른 장르의 언급이 나타나 웹툰에 대한 독자의 불만족 요인을 파악할 수 있었다.

• Journal of the Korea Convergence Society
• /
• v.10 no.7
• /
• pp.109-114
• /
• 2019

This study explored how the tendency of writing 'comment' by Facebook users and the number of 'like' in posting message affected to product attitude, purchasing intention. One hundred thirty five male and female college students were divided into groups with high/low tendency of writing 'comment'. The subjects had to read posting message about athlete shoes on Facebook's newsfeed, different from the conditions under which the 'like' in the posting was high and low. Then, they were responded product attitude and the intention of purchasing. The results of two-way ANOVA showed that the users with low tendency of writing 'comment' displayed more positive product attitude and higher willingness to purchase under condition with a high 'like' number of posting than under condition with a low 'like' number of it.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.27 no.4
• /
• pp.63-84
• /
• 2016

This study aimed to investigate the extent that users participated in Facebook of U.S. state archives and the types of user responses to posts on the Facebook. For the purpose, data created between August 1st and September 30th in 2016 were collected from Facebook continuously operated by 27 state archives. The extent of user participation was measured based on the number of user comments, the number of unique commenters, and the average number of comments per post. According to the measures, top 10 Facebook of state archives were selected. Out of these, Facebook of Ohio (1st), Florida (5th) and Arkansas (10th) state archives were chosen to collect 687 user comments and 132 posts. The analysis showed that comments regarding users' emotional opinion and judgement, adding explanations to a post, and sharing personal stories occupied a large portion. Interactions among users or between a user and an archivist were also identified. With regard to posts, those for sharing information/knowledge of records held in archives were identified as a high percentage. The study suggested that archives should collect and present historical information and related records connected to users' lives, examine methods for effective communication with users via social media and facilitate publicity and outreach services of archives based on shaping and maintaining online user community through social media.

• The Journal of the Korea Contents Association
• /
• v.12 no.2
• /
• pp.40-51
• /
• 2012

Recently, many users discuss and argue with others using replying comments. This implies that a series of comments can be a new source of information since various opinions can be appeared in the dispute. It is important to understand the implicit dispute structure immanent in the comment set. In this paper, we examine the characteristics of disputes using replying comments in the Internet forum sites using a set of test articles with the comments collected from SketicalLeft and Agora, which are famous Internet forum sites in Korea. And we propose a new method for detecting and visualizing the dispute sections and relations from a large set of replying comments. To show the performance of our method, we measured precision, recall, and F-measure. According to the experimental results, the F-measures of the detection of the comments in dispute are about 0.84 (SketpcialLeft) and 0.83 (Agora); those of the detection of the commenter pairs in dispute are 0.75 (SketpcialLeft) and 0.82 (Agora), respectively. Since our method exploits the temporal order of commenters to detect the disputes, it is not dependent on the host language nor on the typos in comments. Also, our method can help the readers to grasp the structure of controversy hidden in the comment set through the visualized view.