• Title/Summary/Keyword: News Article

Search Result 233, Processing Time 0.024 seconds

Semi-supervised GPT2 for News Article Recommendation with Curriculum Learning (준 지도 학습과 커리큘럼 학습을 이용한 유사 기사 추천 모델)

  • Seo, Jaehyung;Oh, Dongsuk;Eo, Sugyeong;Park, Sungjin;Lim, Heuiseok
    • Annual Conference on Human and Language Technology
    • /
    • 2020.10a
    • /
    • pp.495-500
    • /
    • 2020
  • 뉴스 기사는 반드시 객관적이고 넓은 시각으로 정보를 전달하지 않는다. 따라서 뉴스 기사를 기존의 추천 시스템과 같이 개인의 관심사나 사적 정보를 바탕으로 선별적으로 추천하는 것은 바람직하지 않다. 본 논문에서는 최대한 객관적으로 다양한 시각에서 비슷한 사건과 인물에 대해서 판단할 수 있도록 유사도 기반의 기사 추천 모델을 제시한다. 길이가 긴 문서 사이의 유사도를 측정하기 위해 GPT2 [1]언어 모델을 활용했다. 이 과정에서 단방향 디코더 모델인 GPT2 [1]의 단점을 추가 학습으로 개선했으며, 저장 공간의 효율과 핵심 문단 추출을 위해 BM25 [2]함수를 사용했다. 그리고 준 지도 학습 [3]을 통해 유사도 레이블링이 되어있지 않은 최신 뉴스 기사에 대해서도 자가 학습을 진행했으며, 이와 함께 길이가 긴 문단에 대해서도 효과적으로 학습할 수 있도록 문장 길이를 기준으로 3개의 단계로 나누어진 커리큘럼 학습 [4]방식을 적용했다.

  • PDF

Designing Effective Summary Models for Defense Articles with AI and Evaluating Performance (AI를 이용한 국방 기사의 효과적인 요약 모델 설계 및 성능 평가)

  • Yerin Nam;YunYoung Choi;JongGeun Choi;HyukJin Kwone
    • Journal of the Korean Society of Systems Engineering
    • /
    • v.20 no.1
    • /
    • pp.64-75
    • /
    • 2024
  • With the development of the Internet, the information in our lives has become fast and diverse. Especially in the field of defense, articles and information are pouring in from various sources every day, and fast information selection, understanding, and decision-making are required in the ever-changing situation. It is very cumbersome to go from platform to platform and read articles one by one to get the information you need. To solve this problem, this research aims to save time and provide quick access to the latest information by allowing you to quickly grasp key information from summarized content without having to read the entire article. This can improve efficiency by allowing defense professionals to focus more on important tasks rather than extensive information search and analysis.

Study of major issues and trends facing ports, using big data news: From 1991 to 2020 (뉴스 빅데이터를 활용한 항만이슈 변화연구 : 1991~2020)

  • Yoon, Hee-Young
    • Journal of Korea Port Economic Association
    • /
    • v.37 no.1
    • /
    • pp.159-178
    • /
    • 2021
  • This study analyzed issues and trends related to ports with 86,611 news articles for the 30 years from 1991 to 2020, using BIGKinds, a big data news analysis service. The analysis was based on keyword analysis, word cloud, relationship diagram analysis offered by BIG Kinds. Analysis results of issues and trends on ports for the last 30 years are summarized as follows. First, during Phase 1 (1991-2000), individual ports such as Busan, Incheon, and Gwangyang ports tried to strengthen their own competitiveness. During Phase 2 (2001-2010), efforts were made on gaining more professional and specialized port management abilities by establishing the Busan Port Authority in 2004, the Incheon Port Authority in 2005, and the Ulsan Port Authority in 2007. During Phase 3 (2011-2020), the promotion of future-oriented, eco-friendly, and smart ports was major issues. Efforts to reduce particulate matters and pollutants produced from ports were accelerated, and an attempt to build a smart port driven by port automation and digitalization was also intensified. Lastly, in 2020, when the maritime sector was severely hit by the unexpected shock of the COVID-19 pandemic, a microscopic analysis of trends and issues in 2019 and 2020 was made to look into the impact the pandemic on the maritime industry. It was found that shipping and port industries experienced more drastic changes than ever while trying to prepare for a post-pandemic era as well as promoting future-oriented ports. This study made policy suggestions by analyzing port-related news articles and trends, and it is expected that based on the findings of this research, further studies on enhancing the competitiveness of ports and devising a sustainable development strategy will follow through a comparative analysis of port issues of different countries, thereby making further progress toward academic research on ports.

A Study on the Online Newspaper Archive : Focusing on Domestic and International Case Studies (온라인 신문 아카이브 연구 국내외 구축 사례를 중심으로)

  • Song, Zoo Hyung
    • The Korean Journal of Archival Studies
    • /
    • no.48
    • /
    • pp.93-139
    • /
    • 2016
  • Aside from serving as a body that monitors and criticizes the government through reviews and comments on public issues, newspapers can also form and spread public opinion. Metadata contains certain picture records and, in the case of local newspapers, the former is an important means of obtaining locality. Furthermore, advertising in newspapers and the way of editing in newspapers can be viewed as a representation of the times. For the value of archiving in newspapers when a documentation strategy is established, the newspaper is considered as a top priority that should be collected. A newspaper archive that will handle preservation and management carries huge significance in many ways. Journalists use them to write articles while scholars can use a newspaper archive for academic purposes. Also, the NIE is a type of a practical usage of such an archive. In the digital age, the newspaper archive has an important position because it is located in the core of MAM, which integrates and manages the media asset. With this, there are prospects that an online archive will perform a new role in the production of newspapers and the management of publishing companies. Korea Integrated News Database System (KINDS), an integrated article database, began its service in 1991, whereas Naver operates an online newspaper archive called "News Library." Initially, KINDS received an enthusiastic response, but nowadays, the utilization ratio continues to decrease because of the omission of some major newspapers, such as Chosun Ilbo and JoongAng Ilbo, and the numerous user interface problems it poses. Despite these, however, the system still presents several advantages. For example, it is easy to access freely because there is a set budget for the public, and accessibility to local papers is simple. A national library consistently carries out the digitalization of time-honored newspapers. In addition, individual newspaper companies have also started the service, but it is not enough for such to be labeled an archive. In the United States (US), "Chronicling America"-led by the Library of Congress with funding from the National Endowment for the Humanities-is in the process of digitalizing historic newspapers. The universities of each state and historical association provide funds to their public library for the digitalization of local papers. In the United Kingdom, the British Library is constructing an online newspaper archive called "The British Newspaper Archive," but unlike the one in the US, this service charges a usage fee. The Joint Information Systems Committee has also invested in "The British Newspaper Archive," and its construction is still ongoing. ProQuest Archiver and Gale NewsVault are the representative platforms because of their efficiency and how they have established the standardization of newspapers. Now, it is time to change the way we understand things, and a drastic investment is required to improve the domestic and international online newspaper archive.

An Analysis of Social Perception on Forest Using News Big Data (뉴스 빅데이터를 활용한 산림에 대한 사회적 인식 변화 분석)

  • Jang, Youn-Sun;Lee, Ju-Eun;Na, So-Yeon;Lee, Jeong-Hee;Seo, Jeong-Weon
    • Journal of Korean Society of Forest Science
    • /
    • v.110 no.3
    • /
    • pp.462-477
    • /
    • 2021
  • The purpose of this study was to understand changes in domestic forest policy and social perception of forests from a macro perspective using big data analysis of news articles and editorials. A total of 13,570 'forest' related data were collected from metropolitan and economic journals from 1946-2017 using keyword and CONCOR (Convergence of iterated Correlations) analysis. First, we found the percentage of articles and editorials using the keyword 'forest'increased overall. Second, news data on 'forest' in the field of reporting was concentrated in the "social" sector during the first period (1946-1966), followed by forest-related issues expanding to various fields from the second (1967-1972) to fifth (1988-1997) periods, then toward the "culture" sector in the sixth (1998-2007) and "politics" after the seventh (2008-2017) period. Third, we found changes in the policy paradigm over time significantly changed social awareness. In the first and second periods, people experienced livelihood issues rather than forest greening or forest protection policy and expanded their awareness of planned and scientific afforestation (third) to environmental protection (fourth) and ecological perspectives (sixth to seventh). The key outcome of our analysis was leveraging news big data that reflected polices on forests and public social perception To further derive future social issues,more in-depth analysis of public discourse and perception will be possible using textual big data and GDP of various social network services (SNS), such as combining blogs and YouTube.

Funology Body : Classified Application System Based on Funology and Philosophy of the Human Body (Funology Body : Funology와 '몸의 철학' 이론을 바탕으로 한 어플리케이션 분류 검색 체계 연구)

  • Kihl, Tae-Suk;Jang, Ju-No;Ju, Hyun-Sun;Kwon, Ji-Eun
    • Science of Emotion and Sensibility
    • /
    • v.13 no.4
    • /
    • pp.635-646
    • /
    • 2010
  • This article focuses on Funology and a new classified application system based on concept of language and thought which are formed by body experience. It is defined by Funology Body as that. Funology Body is classifying and searching system which are consisted of a body, world (environment), and device tool. The body is sectioned by Brain, Eyes, Ears, Nose, Mouth, Hand, Torso, Feet, and Heart according as parts of the human body. This allows intuiting and experience searching as making classified system connected to the application relationship with concept of an each part of body. The Brain of the body is sub-classified by Book, Account, Business, Memory, Education, Search, and Aphorism to imply the application with thought. The Eyes take Video, Photography, and Broadcast for visibility. The Ears is categorized as Music, Instrument, Audio, and Radio for hearing. The Nose gets Perfume, Smell for olfactory sense. The Mouth is sectioned by Food, SNS, Chatting, Email, and Blog for eating and communication. The Hand sorts into Games, Kits, and Editing to handle, create, and play. The Torso is grouped by Health, Medical, Dance, Sport, Fashion, and Testyuorself related by protecting internal and meaning of the body core. The Feet is classified by Travel, Transportation, Map, and Outdoor for moving and concept of expanding the terrain. The Heart is consisted of Fear, Anger, Joy, Sadness, Acceptance, Disgust, Expectation, and Surprise for a human feeling. Beyond that, the World takes News, Time, Weather, Map, Fortune, and Shop, and Device tool gets Interface, Utilities. The Funology Body has a unique characteristic of giving intuitive and sensuous pleasure and reflection of users' attitude and taste for changing application flexibly.

  • PDF

Comparing Attitudes to Emergency Grant for all Citizen in the News Articles (전국민재난지원금에 대한 신문기사 보도태도 차이·변화 연구)

  • Bae, Hwa-Sook
    • The Journal of the Korea Contents Association
    • /
    • v.21 no.9
    • /
    • pp.806-816
    • /
    • 2021
  • This study aims to describe the reporting patterns of articles on the Emergency Grant for all citizens(EGAC) and compare the reporting attitudes by newspaper companies. The main results are as follows: First, conservative newspapers reported more than four times as much coverage of EGAC as liberal newspapers. Newspaper articles reported during the week of the National Assembly Election Day accounted for about a quarter of the total, and 79.1% of the articles were reported in the political realm of newspapers, and only 2.8% in the social realm. Second, a conservative newspaper reported a critical attitude toward EGAC at 52.6% and favorable articles at 5.3%. On the other hand, in a liberal newspaper, critical articles were 17.1% and favorable articles were 37.1%. The inefficiency of selectivism was reported as the basis for the argument in favor, and concerns about the burden of deterioration in the financial soundness of the opponents were reported the most. Politicians are the most cited sources of information in articles. Finally, in prior to policymaking, the proportion of the media in favor of and against the news was similar, and the proportion of articles with a neutral attitude accounted for more than half. And in the specific method discussion stage, the articles in favor of the article exceeded the proportion of articles on the opposite side.

Complaint-based Data Demands for Advancement of Environmental Impact Assessment (환경영향평가 고도화를 위한 평가항목별 민원기반 데이터 수요 도출 연구)

  • Choi, Yu-Young;Cho, Hyo-Jin;Hwang, Jin-Hoo;Kim, Yoon-Ji;Lim, No-Ol;Lee, Ji-Yeon;Lee, Jun-Hee;Sung, Min-Jun;Jeon, Seong-Woo;Sung, Hyun-Chan
    • Journal of the Korean Society of Environmental Restoration Technology
    • /
    • v.24 no.6
    • /
    • pp.49-65
    • /
    • 2021
  • Although the Environmental Impact Assessment (EIA) is continuously being advanced, the number of environmental disputes regarding it is still on the rise. In order to supplement this, it is necessary to analyze the accumulated complaint cases. In this study, through the analysis of complaint cases, it is possible to identify matters that need to be improved in the existing EIA stages as well as various damages and conflicts that were not previously considered or predicted. In the process, we dervied 'complaint-based data demands' that should be additionally examined to improve the EIA. To this end, a total of 348 news articles were collected by searching with combinations of 'environmental impact assessment' and a keyword for each of the six assessment groups. As a result of analysis of collected data, a total of 54 complaint-based data demands were suggested. Among those were 15 items including 'impact of changes in seawater flow on water quality' in the category of water environment; 13 items including 'area of green buffer zone' in atmospheric environment; 10 items including 'impact of soundproof wall on wind corridor' in living environment; 8 items including 'expected number of users' in socioeconomic environment, 4 items including 'feasibility assessment of development site in terms of environmental and ecological aspects' in natural ecological environment; and 4 items including 'prediction of sediment runoff and damaged areas according to the increase in intensity and frequency of torrential rain' in land environment. In future research, more systematic complaint collection and analysis as well as specific provision methods regarding stages, subjects, and forms of use should be sought to apply the derived data demands in the actual EIA process. It is expected that this study can serve to advance the prediction and assessment of EIA in the future and to minimize environmental impact as well as social conflict in advance.

Sports Celebrities as a Determinant of Sport Media Distribution Contents: Focusing on Tacit Premise of Agenda Setting Theory (스포츠미디어의 유통 콘텐츠 결정요인으로서 스포츠 스타: 의제설정 이론의 암묵적 전제를 중심으로)

  • YOO, Sang-Keon;KIM, Yong-Eun;SEO, Won-Jae
    • Journal of Distribution Science
    • /
    • v.17 no.10
    • /
    • pp.83-91
    • /
    • 2019
  • Purpose - Media is a significant distributional channel in sport. In terms of determining the influencer in building sport media contents, recent sport media studies have employed agenda-setting theory, assuming media itself as the agenda provider. In a real-world situation, however, sports stars have been deemed key factor determining distribution contents in sport. The starting point of this study is the "tacit premise" of agenda-setting theory. Given the agenda-setting theory, the current study attempted to explore the function of sport stars as an agenda provider, which is a key determinant of sport distribution. Research design, data, and methodology - This study has reviewed articles of Yuna Kim, Sang-hwa Lee, and Hyun-jin Ryu from daily newspapers including as dong-a ilbo and joongang ilbo (2013 to 2017). The study collected data, portable document format (PDF), from the online archive of dong-a ilbo and joongang ilbo. We coded the length of the article, the frequency, the size of the picture, and the structural form of the article. Inter-coder reliability was compared with data previously investigated by the researcher. Inter-coder reliabilities for study 1 and 2 was .89 and .85. To examine hypotheses, descriptive analysis, correlations, and cross-tap analysis were performed. Results - The results partially supported the hypotheses proposing the significant role of sports stars as the agenda setters in distributing sport media contents. In specific, the study found that the number of articles about sports stars prevailed the number of articles about regular athletes. Besides, studies found that the use of photos was more frequent in articles of sports starts than that of regular athletes. In sports newspaper articles, featured story articles were used more than straight-articles for news relating to sports stars. Also, sports newspaper of sports stars contained more information associated within an event rather than outside of an event. Conclusions - In sports journalism, this study challenges the current theory that the media affects the composition and the content of sports coverages. As the principle of the agenda-setting of sports media, the influence of sports stars must be continuously studied along with a follow-up study.

The Characteristic of Media Consumer and Legal Principles for Consumer Movements Protection (언론소비자의 특성과 소비자운동의 보호법리 - 광고불매운동을 중심으로)

  • Lee, Seung-Sun
    • Korean journal of communication and information
    • /
    • v.48
    • /
    • pp.5-24
    • /
    • 2009
  • This study is aimed to analyze the concept of media consumer and legal principles for consumer movements protection. Based on the concept and legal principles, this research is to review the characteristics of the advertisement boycott campaign. Article 124 of the Constitution prescribes that the state should guarantee the consumer protection movements. According to the Article 4 of the Framework Act on Consumer, consumers have the fundamental right to obtain proper compensation for damages sustained due to use of goods and etc. according to prompt and fair procedure. The type of boycott can be classified into two pattern on the basis of boycott's target or object. They are primary boycott and. secondary boycott. Consumer's boycott independent of primary or secondary, are under the protection of the consumer's right. Media consumers use scarce resources to satisfy their wants and needs to acquire news information and advertising information. Their resources are time and money. Therefore, ads boycott campaign or media boycott campaign is the primary boycott. Consumer's right should be guaranteed to the maximum. The Constitution and consumer protection law should protect the practice of consumer's right, especially consumer's boycott campaign.

  • PDF