• 제목/요약/키워드: 정보경영학

검색결과 1,651건 처리시간 0.023초

머신러닝을 이용한 학업중단 위기학생 관리시스템의 설계 (Design of the Management System for Students at Risk of Dropout using Machine Learning)

  • 반재훈;김동현;하종수
    • 한국전자통신학회논문지
    • /
    • 제16권6호
    • /
    • pp.1255-1262
    • /
    • 2021
  • 학업을 중단하는 학생들의 비율이 해마다 증가하고 있어 대학은 학업중단을 막기 위하여 위험요소를 파악하고 이를 사전에 제거하기 위해 노력하고 있다. 그러나 특정 위험요소의 단변수 분석을 통해 위기학생을 관리하고 있어 예측이 부정확한 문제가 발생하고 있다. 본 연구에서는 이러한 문제점을 해결하기 위하여 학업중단 위험요소를 파악하고 학업중단 예측을 위해 머신러닝 방법을 통해 다변수 분석을 실시한다. 또한 다양한 예측방법별로 성능평가를 수행하여 최적화 방법을 도출하고 학업중단을 발생시키는 위험요소간의 연관성과 기여도를 평가한다.

중소기업용 스마트팩토리 보안 취약점 분류체계 개발: 산업제어시스템 중심으로 (Developing a Classification of Vulnerabilities for Smart Factory in SMEs: Focused on Industrial Control Systems)

  • 정재훈;김태성
    • 한국IT서비스학회지
    • /
    • 제21권5호
    • /
    • pp.65-79
    • /
    • 2022
  • The smart factory has spread to small and mid-size enterprises (SMEs) under the leadership of the government. Smart factory consists of a work area, an operation management area, and an industrial control system (ICS) area. However, each site is combined with the IT system for reasons such as the convenience of work. As a result, various breaches could occur due to the weakness of the IT system. This study seeks to discover the items and vulnerabilities that SMEs who have difficulties in information security due to technology limitations, human resources, and budget should first diagnose and check. First, to compare the existing domestic and foreign smart factory vulnerability classification systems and improve the current classification system, the latest smart factory vulnerability information is collected from NVD, CISA, and OWASP. Then, significant keywords are extracted from pre-processing, co-occurrence network analysis is performed, and the relationship between each keyword and vulnerability is discovered. Finally, the improvement points of the classification system are derived by mapping it to the existing classification system. Therefore, configuration and maintenance, communication and network, and software development were the items to be diagnosed and checked first, and vulnerabilities were denial of service (DoS), lack of integrity checking for communications, inadequate authentication, privileges, and access control in software in descending order of importance.

DID와 PSM 분석을 통한 건설업 클린사업장 대상 재정지원사업 평가 (Evaluation of Accident Prevention Financial Support Projects in the Construction Industry Using DID and PSM)

  • 장성은;유성열;김화일
    • 한국산업보건학회지
    • /
    • 제32권1호
    • /
    • pp.31-40
    • /
    • 2022
  • Objectives: The purpose of this study is to understand the impact on accident reduction by analyzing the policy effectiveness of an industrial accident prevention clean workplace support project targeting the construction industry. Methods: In this study, DID and PSM models were used to analyze workplaces receiving and not receiving financial support based on the status of industrial accidents in recent years and the status of the workplaces by year. Results: The research results show that meaningful effects continued to occur in terms of reducing accidents and increasing employment. Conclusions: In this study, we show the effectiveness of the clean workplace support project, one of the South Korean government's financial support projects in the field of occupational safety and health focusing on the construction industry. Financial support projects such as clean projects should be further expanded based on the results of this study.

국제 유가 변동성이 국내 휘발유 가격 비대칭성에 미치는 영향 (An Effect of Volatility of Crude Oil Price on Asymmetry of Domestic Gasoline Price Adjustment)

  • 김남제;김형건
    • 아태비즈니스연구
    • /
    • 제14권1호
    • /
    • pp.351-364
    • /
    • 2023
  • Purpose - This study examines the effects of Dubai oil price and the volatility on the asymmetry of domestic gasoline price adjustment. Additionally, the study investigates the effects of "Altteul" gas-station and tax-cut policies on asymmetry. Design/methodology/approach - Firstly, the study calculates proxies for asymmetry and volatility of each window(every 3-month) by error-correction model and GARCH(1, 1) using daily domestic gas price and Dubai oil price from 2008/04/15 to 2022/12/31. Secondly, the study investigates the effects of the increasing rate of Dubai oil price, volatility, "Altteul" gas-station and tax-cut policies on asymmetry. The autoregressive distributed lag regression model is employed for estimations. Findings - The study finds that changes in the increasing rate of Dubai oil price and both types of volatility of Dubai oil price increase asymmetry. While "Altteul" gas-station and tax-cut policies decrease asymmetry. Additionally, the study fails to find that asymmetry in the Korean gasoline market in the estimation with total observations. Research implications or Originality - An increase in Dubai oil price volatility means an increase in cost uncertainty for gas-station owners. Since cost uncertainty is a kind of financial risk, the increase in volatility reinforces the asymmetry. The study provides supporting evidence for the idea.

4가지 관점의 마이데이터 융합서비스 활성화 전략: 유럽과 한국을 비교하여 (Strategies to Activate MyData Convergence Services from Four Perspectives: Compared to Europe and Korea )

  • 박주석;김혜영;김한성;최민령
    • 한국빅데이터학회지
    • /
    • 제6권2호
    • /
    • pp.181-195
    • /
    • 2021
  • 개인데이터의 활용을 넘어서 마이데이터로의 전환에 대한 관심이 세계적으로 높은 가운데, 실질적으로 마이데이터의 이념이 구현되고 발전하기 위해서는 마이데이터의 사상에 충실한 유용한 서비스가 다양하게 개발되어야 한다. 특히 마이데이터의 발전을 위해서는 무엇보다 융합이 전제되어야 한다. 본 연구는 마이데이터 융합서비스 활성화 전략을 균형 잡힌 네 가지 관점(BLTS: Business, Legal, Technology, Social)에서 제시하였다. 마이데이터 분야에서 가장 선구적이라는 유럽의 정책과 서비스 등을 국내 상황과 비교하여 국내의 융합서비스 활성화 방안을 제시하는 것을 목적으로 하였다. 본 연구를 통해 마이데이터 산업 및 정책에서 균형적이면서 발전적인 아이디어를 얻을 수 있기를 기대한다.

심혈관질환 위험 예측을 위한 비용민감 학습 모델 (Cost-Sensitive Learning for Cardio-Cerebrovascular Disease Risk Prediction)

  • 이유나;이경희;조완섭
    • 한국빅데이터학회지
    • /
    • 제6권2호
    • /
    • pp.161-168
    • /
    • 2021
  • 본 연구에서는 기계 학습을 사용하여 심혈관 질환 예측 모델을 제안한다. 먼저 두 집단간에 다양한 차이를 다차원분석하고 그 결과를 시각화한다. 특히, 질환과 같이 정상집단과 환자집단 간에 높은 클래스 불균형이 존재하는 경우에 대하여 민감도를 향상시킬 수 있는 비용 민감 학습을 사용하는 예측 모델을 제안한다. 본 연구에서는 대표적인 머신러닝 기술인 CART와 XGBoost를 사용하여 예측모델을 개발하고, 심혈관 질환 환자 데이터를 대상으로 예측하고 성능을 비교한다. 연구결과에 따르면 CART가 XGBoost 보다 더 높은 정확도와 특이도를 보였으며, 정확도는 약 70%~74%로 나타났다.

디지털무역 자유화와 개인정보보호의 무역 제한적 요소에 대한 연구 : EU GDPR을 중심으로 (A Study on the Liberalization of Digital Trade and Trade Restrictiveness Factors of Data Privacy : Focusing on EU GDPR)

  • 우기훈;신성식
    • 무역학회지
    • /
    • 제45권3호
    • /
    • pp.71-89
    • /
    • 2020
  • This study was carried out to identify the impact of EU GDPR on international trade amid the ongoing digital trade liberalization. To do this, we first looked at the current trend of digital trade liberalization, the role of data in it, and the trade-restrictive elements of EU GDPR. This allowed us to identify the negative impact of GDPR on free trade. It then conducted an interview survey on Korean companies operating in the EU to verify the conclusions reached. The result of this survey showed that the level of GDPR risk perceived by Korean firms was very low compared with those of American, Japanese and Chinese firms. In particular, the impact of GDPR is not clear for Korea's SMEs. It can be assumed that the reason for this is that Korean SMEs are not using data as a major business tool while the capability of SMEs is sufficient to cope with GDPR. In this regard, the government's appropriate policies and further research for SMEs are needed.

해상보험증권의 해석상 작성자 불이익의 원칙의 적용에 관한 연구 (A study on the Application of the Contra Proferentem Rule in the Interpretation of Marine Insurance Policies)

  • 김성후;한낙현
    • 무역학회지
    • /
    • 제45권5호
    • /
    • pp.279-301
    • /
    • 2020
  • In the absence of any guidance under statutory law, such as the Rules for Construction of Policy, MIA 1906, judges should follow the general principles of interpretation that apply to all contracts. In simple terms, Contra Proferentem Rule means that if the contents of the terms and conditions are ambiguous, they are interpreted against the writer of the terms and conditions. In the Anglo-American Contract Law, the 'default rule' is an important judicial tool that can supplement defects in contract norms and reinforce the principle of private autonomy through gap-filling techniques related to the interpretation of contracts. In Korea, it is sometimes mentioned in case of precedent, and it has been established as a clear rule. This study analyzes the interpretation of terms and conditions is not in the form that the interpretation of other general contracts and other interpretation principles are valid, but contracts based on terms and conditions are also contracts, and as a general rule, the interpretation of terms and conditions is explained like the general contract interpretation.

Probit 회귀분석을 통한 ISMS 인증 취득 전략 탐색: 조직 특성을 중심으로 (Exploring the Strategy for Acquiring ISMS Certification through Probit Regression: Focusing on Organizational Characteristics)

  • 김선주;김태성
    • 한국IT서비스학회지
    • /
    • 제23권1호
    • /
    • pp.11-25
    • /
    • 2024
  • In the field of information security management systems, one of the representative certifications in Korea is ISMS-P certification, and internationally, ISO/IEC 27001 certification is recognized. When companies acquire both ISMS-P (or ISMS) and ISO/IEC 27001 certifications, budget and manpower are duplicated in similar areas. Therefore, it is necessary for the company to choose and invest in a certification that is suitable for its conditions. This paper proposes a strategy for obtaining information security management system certification that is suitable for the characteristics of the company, allowing for effective information security management based on the company's conditions. To achieve this, data were collected from the Ministry of Science and ICT's Information Security Disclosure System (ISDS), the Korea Internet & Security Agency (KISA), and the Financial Supervisory Service's Data Analysis, Retrieval and Transfer System (DART), and Probit regression analysis was conducted. During the Probit regression analysis, the relationships between seven independent variables and five cases of ISMS-P (or ISMS) acquisition, ISMS-P acquisition, ISMS acquisition, ISO/IEC 27001 acquisition, and both ISMS-P (or ISMS) and ISO/IEC 27001 acquisition were analyzed. The analysis results revealed the relationship between company characteristics, including industry, and certification acquisition in the ISMS field. Through this, strategies for certification acquisition based on company types could be suggested.

내부통제와 미래이익에 대한 주가 정보성 (Internal Control and Stock Price Informativeness about Future Earnings)

  • 왕람;박희우
    • 아태비즈니스연구
    • /
    • 제14권4호
    • /
    • pp.255-273
    • /
    • 2023
  • Purpose - This study examines whether internal control has an effect on stock price informativeness about future earnings. High quality internal control provides continuous assurance for the quality of financial reports, and these future earnings-related information is accurately reflected in the current stock price. Design/methodology/approach - This study collected 12,862 data from 2006 to 2021 in China to make an empirical analysis using the future earnings response coefficient (FERC) and the multiple regression analysis were hired in order to analyze the data. Findings - We find that internal control strengthens the association between current returns and future earnings, indicating that more information about future earnings is reflected in current stock prices. This positive effect exists in both the main board market and the growth enterprise market of China's stock market, especially in the main board market after the implementation of the internal control policy. In addition, we find that the positive effect is weaker for firms that report internal control deficiencies or receives non unqualified internal control audit opinions. The results using earnings persistence yield similar findings, further supporting the results based on the FERC model. Research Implications or Originality - Our tests provide strong evidence that the quality of internal control affects FERC in China stock market.