자연어 처리를 이용한 유튜브 댓글 여론 변화 분석

박진우,서하영,이선호,은지우,김익수 한국지식정보기술학회 2022 한국지식정보기술학회 논문지 Vol.17 No.4

As information in online media has increased, various studies have been conducted on the effect of comments on public opinion. However, in the existing studies, experiments and surveys were conducted on individuals. This cannot be considered appropriate in that the situation of writing comments in an actual online environment is anonymous. In this paper, in order to overcome the limitations of existing research papers, we chose a method of selecting a medium in which public opinion has hardened to some extent and observing it. Among many online media, we selected YouTube, in which opinion sharing is most actively performed recently and information on comments is easy to obtain and specified the videos on topics in which public opinion could be changed over time. After that, we crawled comments on the videos and measured the degree of the emotion of the comments by conducting a natural language processing-based sentiment analysis using the KcELECTRA model. In addition, we visualized this and analyzed how much public opinion of the latter period is affected by early public opinion. The analysis result of this paper show that the probability of public opinion reversal is very low, which means that the probability that the initial public opinion will become the final public opinion is very high. Based on the conclusion of this study, we expect the public to realize that the public can be one of the users who are swept by public opinion, and to accept information from a more critical and objective perspective with caution.

Online Unstructured Data Analysis Models with KoBERT and Word2vec: A Study on Sentiment Analysis of Public Opinion in Korean

Changwon Baek,Jiho Kang,SangSoo Choi 한국지능시스템학회 2023 INTERNATIONAL JOURNAL of FUZZY LOGIC and INTELLIGE Vol.23 No.3

Online news articles and comments play a vital role in shaping public opinion. Numerous studies have conducted online opinion analyses using these as raw data. Bidirectional encoder representations from transformer (BERT)-based sentiment analysis of public opinion have recently attracted significant attention. However, owing to its limited linguistic versatility and low accuracy in domains with insufficient learning data, the application of BERT to Korean is challenging. Conventional public opinion analysis focuses on term frequency; hence, low-frequency words are likely to be excluded because their importance is underestimated. This study aimed to address these issues and facilitate the analysis of public opinion regarding Korean news articles and comments. We propose a method for analyzing public opinion using word2vec to increase the word-frequency-centered analytical limit in conjunction withKoBERT, which is optimized for Korean language by improving BERT. Naver news articles and comments were analyzed using a sentiment classification model developed on the KoBERT framework. The experiment demonstrated a sentiment classification accuracy of over 90%. Thus, it yields faster and more precise results than conventional methods. Words with a low frequency of occurrence, but high relevance, can be identified using word2vec.

소셜 네트워크 서비스(SNS: Social Network Service)상의 담론 분석을 통한 인지적,정서적 측면의 여론 변화 연구

홍주현 ( Ju Hyun Hong ) 한국커뮤니케이션학회 2011 커뮤니케이션학연구 Vol.19 No.3

이 연구는 소셜 네트워크 서비스인 트위터에서 단기간에 형성되는 여론을 효율적으로 감지하고, 여론의 변화 시점을 파악하기 위한 새로운 방법을 제안하였다. 찬반으로 나타나는 기존의 지배 여론 감지와 달리 이 연구는 인지적 측면뿐만 아니라 정서적 측면을 고려하여 여론을 분석하였다. 인지적 측면에서는 이슈에 대한 찬반 여론을 분석하였고, 정서적 측면에서는 슬픔, 분노, 노여움 등의 정서가 트윗에 어떻게 나타나는지 분석하였다. 또한, 시시각각 변화하는 트윗 여론을 인지적 측면과 정서적 측면에서 파악하였다. 인지적 측면에서는 시기별로 주요 단어를 추출함으로써 시간의 경과에 따른 여론의 변화를 감지하였다. 정서적 측면에서는 트윗 문장을 긍정적 트윗과 부정적 트윗으로 구분하여 측정한 후 그 차이를 구하여 시기별 ``정서적 주기성``을 분석하였다. 각 주제별 ``정서적 주기성``을 그래프로 나타내 변동성이 큰 트위터 여론을 한눈에 파악할 수 있었다. 이 연구는 텔레비전토론 직후 트위터에 개진된 의견을 분석함으로써 텔레비전 토론에서 논의된 주제에 대한 실시간 여론을 효율적으로 감지할 수 있는 이론적 관점을 제공하였다. 인지적?정서적 측면에서 트위터 여론의 변화를 분석한 이 연구의 결과는 정부의 정책 수립과 정이나 선거와 같은 실무적 측면에서 유용하게 활용될 수 있을 것이다. This study aims to explore the detection of public opinion and the periodicity of sentiment measurable in aggregated Twitter opinion. For detecting Twitter opinion, this study measured user attitude to certain issue as negative, positive, mixed and other. For characterizing the debate, whether Twitter users favored a particular theme or not, this study analyzed overall tweets by text analysis and then selected main themes according to time flow. This study also focuses on the periodicity of sentiment observable in the tweets and charted the change of opinion on a graph. Finally, this study showed the relationship between main theme of tweet during the TV debate and the sentiment of tweet. This will enable them to better understand the temporal cycle of sentiment in Twitter. In conclusion, the results of this study is useful for analysing public opinion via Twitter coupled with television debate about controversial issues. It offers the volatility of public opinion and sentiment to television debate. Considering the main theme and sentiment according to time flow in analysing public opinion will help public officers, political candidates detect public opinion via Twitter.

교육정책관련 여론탐색을 위한 소셜미디어 감정분석 연구

정진명,유기영,구찬동 한국정보화진흥원 2017 정보화정책 Vol.24 No.4

With the development of social media services in the era of Web 2.0, the public opinion formation site has been partially shifted from the traditional mass media to social media. This phenomenon is continuing to expand, and public opinions on government polices created and shared on social media are attracting more attention. It is particularly important to grasp public opinions in policy formulation because setting up educational policies involves a variety of stakeholders and conflicts. The purpose of this study is to explore public opinions about education-related policies through an empirical analysis of social media documents on education policies using opinion mining techniques. For this purpose, we collected the education policy-related documents by keyword, which were produced by users through the social media service, tokenized and extracted sentimental qualities of the documents, and scored the qualities using sentiment dictionaries to find out public preferences for specific education policies. As a result, a lot of negative public opinions were found regarding the smart education policies that use the keywords of digital textbooks and e-learning; while the software education policies using coding education and computer thinking as the keywords had more positive opinions. In addition, the general policies having the keywords of free school terms and creative personality education showed more negative public opinions. As much as 20% of the documents were unable to extract sentiments from, signifying that there are still a certain share of blog posts or tweets that do not reflect the writers' opinions. 웹2.0시대와 더불어 소셜미디어 서비스의 발달로 전통적인 여론형성의 기능이 매스미디어에서 소셜미디어로 일부 이동되었으며, 이런 현상은 계속 확대되고 있어, 정부 정책에 대한 소셜미디어 기반의 여론이 관심을 받고 있다. 특히, 교육정책은 다양한 이해관계자들이 존재하고, 정책의제 설정과정에서도 많은 의견의 충돌이 발생되기 때문에 정책을 수립함에 있어 대중의 여론을 파악하는 것이 더욱 중요하다고 할 수 있다. 본 연구는 교육정책관련 소셜미디어 서비스를 통해 작성된 문서들을 오피니언 마이닝 기법으로 분석하여, 교육관련 정책에 대한 대중의 여론탐색을 목적으로 하였다. 이를 위하여 소셜미디어 서비스를 통해 사용자들이 생산하는 교육정책 관련 문서들을 키워드 기반으로 수집하고, 토큰화 시킨 후 감성자질을 추출하고 감성사전으로부터 해당 문서의 감정을 점수화 하여, 특정 교육정책 키워드에 대한 대중의 여론을 탐색하였다. 그 결과 디지털교과서, 이러닝 등을 키워드로 하는 스마트교육 정책에 대해서는 긍정보다 부정적인 감정이 많은 것으로 나타났으며, 코딩교육, 컴퓨터적 사고 등을 키워드로 하는 소프트웨어 교육 정책에 대해서는 긍정적인 여론의 방향으로 나타났다. 자유학기제, 창의 인성교육 등을 키워드로 하는 일반 정책에 대해서는 부정적인 여론이 많은 것으로 나타났다. 또한 전체 분석 대상 문서 중에서 감정 자질이 전혀 추출되지 않은 문서가 20%나 되어 블로그나 트위터의 내용에 저자의 의견이 반영되지 않은 내용이 아직 일정 비율로 존재함을 알 수 있었다.

소셜미디어에서의 여론 형성기제에 관한 탐색적 연구 - 20대 대선 ‘여성가족부 폐지’ 빅데이터 분석

감현주,한혜경 부산울산경남언론학회 2023 지역과 커뮤니케이션 Vol.27 No.3

This study sought to discuss the reality of individual opinion expression and public opinion polarization by applying Spiral of Silence Theory to the current social media environment, and to examine how the pattern appears through big data analysis. To this end, social media such as communities, Twitter(current "X"), blogs, and Instagram conducted frequency analysis, association analysis, and sentiment analysis under the keyword "abolition of the Ministry of Gender Equality and Family." As a result, it was found that in today's social media environment, which pursues homogeneous orientation, recommended algorithms, and functions as a reference group, individual opinions and emotions become more active and solid, and over time, public opinion develops into confrontation and conflict. In particular, the use of emotional and negative words increased rapidly after the election was confirmed, and the affective polarization phenomenon was confirmed, in which antipathy between the two camps was maximized. It was observed that individuals' political opinion expression and the polarization of public opinion appeared more clearly on closed platforms than on open platforms, and that a minority of strong opinions involved in political issues heightened emotions in polarized situations through retweets. This study is an exploratory study to incorporate the public opinion-forming mechanism of silence into the actual context of social media today through big data analysis, and aims to contribute to enhancing the value of empirical research and expanding theoretical discussions.

중국 웨이보를 통한 인터넷 민족주의의 확산 연구: 메시지의 확산 추이와 메시지에 대한 정서 분석을 중심으로

홍주현 ( Ju Hyun Hong ),윤해진 ( Hae Jin Yun ) 한국커뮤니케이션학회 2014 커뮤니케이션학연구 Vol.22 No.4

이 연구는 중국의 민족주의가 SNSs를 통해 어떻게 확산되는지 확산 추이를 파동을 중심으로 알아보고, 정서 분석을 통해 여론을 파악했다. 중국인 선장의 한국 해경 살해 사건을 중심으로 웨이보에 올라온 2011개의 의견을 분석 대상으로 했다. SNSs를 통해 여론이 얼마나 역동적으로 형성되는지 파악하기 위해 시간의 경과에 따른 파동을 유형화했다. 시시각각 변화하는 여론을 파악하기 위해서 여론의 주기성 분석을 했다. 여론의 주기성은 메시지에 언급된 긍정적인단어와 부정적인 단어의 차이로 측정했다. 이 연구는 역동적이고 변동성이 큰 SNSs 여론을 파악하기 위해 시간의 경과에 따른 추이 분석을 통해 파동의 유형을 찾고, 주요 쟁점 별로 주기성을 측정하는 새로운 방법을 제시했다는데 연구의 의의가 있다. 연구 결과는 정책 결정자나 저널리스트들이 현안을 단기간에 파악하는데 유용하게 활용될 수 있을 것이다. This study seeks to inquire into the formation of public opinion via socialnetworking sites (SNSs) under the topic “Internet Nationalism.” The researchersselected the 2011 case of a corporal in the Korean Coast Guard killed by a Chineseship captain. The present study focuses on (a) how the messages posted on Weibodiffuses as time passes; and (b) how the propensity of said messages changes withtime. This study highlights the dynamics of public opinion utilizing social wave andthe periodicity of public opinion on SNSs based on sentiment analysis. The maintopic of the issue is then further classified in detail by subject through thecategorization of key words. This method of analysis was used to show the dynamicsand the general leaning of public opinion. Wave type analysis showed public opinionfluctuation while the sentiment analysis showed public opinion changes resulting ineffective detection of public opinion. The present research for the detection ofpublic opinion on SNSs will help policy makers and journalists to gauge publicopinion on controversial issues, giving them the chance to offer alternativeviewpoints.

정책집행 이후 여론과 정책위상 변화에 관한 연구 : 서울시 중앙버스차선제 사례를 중심으로

윤종현,박승규 한국거버넌스학회 2008 한국거버넌스학회보 Vol.15 No.2

This study shed light on uncovering practically about how the public opinion and policy status change and what kind of process they have. For explaining these goals, this research gathered news information about issue from portal site, DAUM, abd took theoretical and statistical analysis such as frequency and cross analysis including econometric analysis, autoregressive conditional heteroskedasticity analysis. According to the results, critically important issue like a change from opposite side to supporting side depend on showing backing of credit worthiness of institution and person. A expression of supporting is possible after verifying visible political outcome, and is prepared complementary counterplan to already executed policy. This research shows that there are short-term through long-term treatment in political changes to cope with public opinion. Government use public relations, incentive, and compensation as tools when it faces severe opposite public opinion. Things to be considered when developing and implementing policy are, firstly, it is needed to understand citizens' general behavior toward policy development and implementation. Secondly, it is needed to heavily focus on policy process. That is, even though the contents of policy are made well by reasonably and scientifically, it need to complement exited policy with acceptance of public opinion. This will enhance political actuality, undoubtedly, adaptation. 본 연구는 정책집행 이후 여론과 정책위상이 어떻게 변하는지를 실증적으로 살펴보고자 하였다. 이를 위하여 ‘서울시 중앙버스차선제’를 중심으로 관련 뉴스 자료를 포털사이트에서 수집하고, 수집된 자료에 대한 내용분석을 실시하였다. 또한 내용분석 과정에서 자료를 계량화 하여 통계분석(빈도분석, 교차분석 그리고 자기회귀조건부이분산분석: autoregressive conditional heteroskedasticity)을 수행하였다. 분석결과에 따르면, 반대 여론에서 지지 여론으로 전환되는 즉, 여론과 정책위상의 변화에는 공신력 있는 기관이나 사람에 의한 지지 표명 그리고 정책적 보완책 마련이 주된 역할을 하고 있었다. 그리고 이러한 지지 표명은 정책의 가시적 성과를 확인할 수 있는 정책집행 후 일정 기간이 지난 후에 이루어졌다. 서울시의 중앙버스차선제에 대한 보완책으로서 단기적 조치로 지하철 정기권 도입과 중 · 장기적 조치인 저공해 버스 도입과 종합교통관리센터 운영 등이 있었다. 특히, 반대 여론에 대한 서울시의 대응은 적극적인 홍보(사과, 대책마련, 버스 환승의 혜택), 유인과 보상(지하철 정기권 도입 등) 등을 사용하였다. 정책의 개발 · 집행 시 고려해야 할 사항으로서 먼저, 정책을 개발하거나 집행함에 있어서 일반시민들의 기존 행태를 충분히 이해하여, 정책집행 시 정책대상자들의 행태 변화와 관련된 거래 비용을 최소화해야 할 것이다. 다음으로, 정책 과정에 보다 충실해야 한다. 즉, 정책 내용이 합리적 · 과학적 방식에 의하여 잘 만들어졌다고 할지라도 이는 극히 가설적이기 때문에 이 정책과 관련된 국민과 일반시민의 견해를 수용해서 정책적 실현가능성을 높이고, 이를 통한 정책대상자들의 정책에 대한 절차적 · 결과적 순응 확보가 이루어져야 할 것이다.

대통령지지도와 언론의 관계를 통해 살펴본 여론의 순환적 형성에 관한 연구

홍원식(Won-Sik Hong) 한국언론학회 2007 한국언론학보 Vol.51 No.6

The goal of this study is to explore the influence of the news media on the cyclical formation process of public opinion with the case of presidential approval ratings. To explore this relationship empirically, this study investigates the historical fluctuations of presidential approval ratings, which have been regularly measured since the 1950s, and how the New York Times' presentation of the ratings has itself influenced public opinion on presidential popularity for the last forty years. To investigate the relationship between the NYT and a series of presidential approval rates in two-ways, this study suggests two hypotheses of exploring the influence of convergence and change in presidential approval ratings on the NYT and two hypotheses of testing the additive/subtractive effect and volatility effect of the NYT re-ports on the following approval rating. The analysis of the relationship is conducted by using two statistical methods: First, in order to test the two hypotheses of investigating the influence of convergence and change in presidential approval rates on the NYT's decision to report the approval rates, this study applies the quadratic regression model. Next, the influence of the NYT reports on the subsequent approval rates is investigated by applying time-series models. The statistical tests confirm that there exists a reciprocal relation between the NYT reports and public evaluation of presidential performance, and the relation results in the intensification of majority opinion. The findings of this study show: first, media sensitivity to a given issue varies over time: the convergence and change in public opinion raise the salience of the given issue; second, news selection does matter: news media provide unambiguous information but they are not an exact sample of the population; third, a simple cue makes a difference in individual statistical sense of the climate of public opinion. This study concludes that public opinion influence me-dia attention, which, in tum, affect public opinion at a later date.

Analyzing Public Opinion with Social Media Data during Election Periods: A Selective Literature Review

Jin-ah Kwak,조성겸 충남대학교 아시아여론연구소 2018 Asian journal for public opinion research Vol.5 No.4

There have been many studies that applied a data-driven analysis method to social media data, and some have even argued that this method can replace traditional polls. However, some other studies show contradictory results. There seems to be no consensus as to the methodology of data collection and analysis. But as social media-based election research continues and the data collection and analysis methodology keep developing, we need to review the key points of the controversy and to identify ways to go forward. Although some previous studies have reviewed the strengths and weaknesses of the social media-based election studies, they focused on predictive performance and did not adequately address other studies that utilized social media to address other issues related with public opinion during elections, such as public agenda or information diffusion. This paper tries to find out what information we can get by utilizing social media data and what limitations social media data has. Also, we review the various attempts to overcome these limitations. Finally, we suggest how we can best utilize social media data in understanding public opinion during elections

Analyzing Public Opinion with Social Media Data during Election Periods: A Selective Literature Review

Kwak, Jin-ah,Cho, Sung Kyum Center for Asian Public Opinion ResearchCollaborat 2018 Asian journal for public opinion research Vol.5 No.4

There have been many studies that applied a data-driven analysis method to social media data, and some have even argued that this method can replace traditional polls. However, some other studies show contradictory results. There seems to be no consensus as to the methodology of data collection and analysis. But as social media-based election research continues and the data collection and analysis methodology keep developing, we need to review the key points of the controversy and to identify ways to go forward. Although some previous studies have reviewed the strengths and weaknesses of the social media-based election studies, they focused on predictive performance and did not adequately address other studies that utilized social media to address other issues related with public opinion during elections, such as public agenda or information diffusion. This paper tries to find out what information we can get by utilizing social media data and what limitations social media data has. Also, we review the various attempts to overcome these limitations. Finally, we suggest how we can best utilize social media data in understanding public opinion during elections.