• Journal of Korea Multimedia Society
• /
• v.24 no.11
• /
• pp.1508-1517
• /
• 2021

In CCTV environments, a lot of learning image data is required to monitor illegal dumping of garbage with a typical image-based object detection using deep learning method. In this paper, we propose a system to monitor unauthorized dumping of garbage by learning the articular points of the person using only a small number of images without immediate use of the image for deep learning. In experiment, the proposed system showed 74.97% of garbage dumping detection performance with only a relatively small amount of image data in CCTV environments.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.07a
• /
• pp.533-534
• /
• 2021

최근 소셜 네트워크 서비스(SNS)를 통한 건강기능식품 과대광고 적발이 증가하면서 SNS를 통해 브랜드를 선택함에 있어 신뢰도가 소비자에게 중요한 요소가 된다. 본 논문에서는 인스타그램의 해시태그를 이용해 게시글을 크롤링 하여 수집된 게시물 데이터를 가공 및 분석한다. 불용어 사전을 구축해 불용어를 제거해준 뒤 브랜드 추출을 진행하고, 건강기능식품 브랜드 5개에 대한 게시글 데이터를 수집한다. 5개 브랜드의 신뢰도 측정을 위해 게시글, 해시태그, 계정명을 분석기준으로 삼아 라벨링 처리를 한다. 라벨링 된 열을 통해 절대적 수치로 점수를 부여하여 백분율로 점수를 표현한다. 신뢰도 점수와 더불어 브랜드의 고객 참여도 건수를 같이 명시해 준다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2020.07a
• /
• pp.505-508
• /
• 2020

본 연구는 트롤링의 유형 중 하나인 그리핑을 목격해 본 경험이 있는 플레이어를 중심으로 설문을 진행해, 사회 심리 현상과 함께 결과를 살펴보고 대처방안에 대한 개선안을 제시하고자 한 것이다. 그리핑은 트롤링의 유형 중 하나로 고의로 게임을 패배하게 만드는 행위이다. 그리핑의 원인으로 게임 내 동조현상이 발생하지 않는다는 점과 복수의 딜레마가 부재하고, 게임의 시스템이 매슬로우의 욕구단계 중 안전욕구를 충족하지 못한다는 점이 해당할 수 있다. 그리핑을 예방하기 위해서는 트롤링의 유형에 대해 개발사 측에서 명칭과 의미를 정립하고 구분 기준과 처벌 수위를 체계적으로 관리할 필요가 있다. 또한, 적발된 그리퍼에게 확실한 처벌을 내려 그리핑을 반복할 수 없도록 해야 함을 제안한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.703-705
• /
• 2022

코로나19 대유행으로 비대면 교육이 보편화되어 온라인 학습과 시험이 교육기관에서 일반화되고 있다. 이러한 급격한 변화로 교육의 공정성 문제와 온라인 시험의 부정행위 문제가 대두되고 있다. 온라인 시험은 대면 시험과는 달리 시험 감독관이 부정행위를 적발하기 어렵기 때문에 응시자의 다양한 환경을 고려하여 정확하게 부정행위를 판별하는 방법이 필요하다. 본 연구에서는 온라인 시험환경에서 응시자의 행동 데이터와 영상데이터를 분석하여 부정행위를 감독관에게 추천하는 시스템을 제안한다. 제안 시스템의 구현을 통해 온라인 시험 환경에서 부정행위를 탐지 기능을 확인한다.

• Korean Journal of Veterinary Research
• /
• v.1 no.1
• /
• pp.53-53
• /
• 1961

During the period from 1956 to 1959, a serological survey was conducted in au effort to assessing the incidence of brucellosis in various domestic animals in Korea. The following results were obtained. 1. Seventy-six dairy cattle which were imported in the year of 1956 had 9 positive reactors (11.84%). Of 1127 goats imported in 1956 and 1958, eleven(0.98%) were reactor animals. 2. In 1956, of 108 dairy cattle tested 10(9.25%) were reactors. In 1937 and 1958, a total of 155 dairy cattle examined hid ten (6.45%) positive reactors. In 1959, of 127 dairy cattle examined two were reactor animals. 3. In 1958 and 1959, 432 goats in some districts in Korea were subject to test for brucellosis and the test revealed five (1.15%) reactors. In 1958, 683 swine serum samples were examined and eleven (1.04%) were positive. In 1959, 1133 Korean cattle were tested and seven samples (0.64%) showed positive reactions. 4. In the month of July in 1939, 580 beef cattle in Cheju National Ranch and Branch were examined and 111 (19.13%) were found to be reactors. In October of the same year, 157 cattle, consisting of reactor and suspicious cattle groups, were tested, of which 71 samples reacted positive and 47 remained suspicious. 5. In December of the year of 1958, there occurred. an outbreak of brucellosis in a swine herd in the Sachon Branch Experimental Station. Seven serological tests on 438 swine serum samples were conducted, of which 122 (27.87%) were found positive. 6. Dairy cattle No. 33 and No. 35 which had been imported in 1956 and detected as highly positive, were examined for a prolonged period to follow the variation of antibody titers. A marked drop in antibody titer was seen two months after the initial test while the re-increase in titer was observer our months after the first examination.

• Journal of Korean Society of Environmental Engineers
• /
• v.31 no.1
• /
• pp.42-50
• /
• 2009

Samples of subway dust were collected by the air filter system of 30 subway stations on Daegu subway line 1 in January 2008. Samples were sieved below 100 ${\mu}m$, and 14 elements were analyzed using ICP after acid extraction. Results obtained from the source assessment of trace elements using enrichment factor showed that Ca, Fe, K, Mg, Mn, Na, V were influenced by natural sources such as weathered rock and resuspended soil, while Cd, Cr, Cu, Ni, Pb and Zn were influenced by anthropogenic sources such as fuel combustion and waste incineration. Concentrations were remarkably higher in components from natural sources than in components from anthropogenic sources. Anthropogenic sources were significantly affected by indoor dusts than outdoor dusts. Results of pollution indices of heavy metals indicated that indoor dusts were more contaminated with heavy metal ions than outdoor dusts. The correlation analysis among trace elements indicated that components were much correlated in the order of natural sources-anthropogenic sources, anthropogenic sources-anthropogenic sources, natural sources-natural sources in both indoor and outdoor dusts. Trace element components of outdoor dusts were largely correlated than those of indoor dusts. In addition, indoor dusts were significantly affected by outdoor dusts rather than depth from the surface or the average daily number of subway passengers.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.18 no.1
• /
• pp.173-188
• /
• 2023

This paper analyzed the Earnings management of firms selected as preliminary Unicorn. If a manager is selected as a preliminary unicorn firm, he can receive financial support of up to 20 billion won, creating a factor in managing the manager's earnings. The motive for management's earnings management is related to the capital market. Accounting information is used by investors and financial analysts, and corporate profits affect corporate value. Therefore, if the accounting earning is adjusted upward, the corporate value will be raised and investment conditions will be favorable. In this paper, earnings quality was measured by the modified Jones model of Dechow et al.(1995) by the ROA control model of Kothari et al.(2005) among the discretionary accruals estimated using an alternative accrual prediction model. Competing similar companies in the same market as the selected companies were formed, and the discretionary accruals were mutually compared to verify the research hypotheses, and only the selected companies were analyzed for the audit year and after the audit year. As a result of the analysis, it was found that the companies selected as preliminary unicorns had higher earnings management compared to the corresponding companies in question, which had a negative impact on the quality of accounting profits. It was found that the companies selected as preliminary unicorns continued to receive incentives for management's earnings management even after being selected. These results indicate that the companies selected as prospective unicorns are recognized for their value in the market through external growth rather than internal growth, and thus, incentives for management's earnings management to attract investment from external investors under favorable conditions are continuing. In the future preliminary unicorn selection evaluation, it was possible to present what needs to be reviewed on the quality of accounting earning. The implication of this paper is that the factors of management's earnings management eventually hinder investors and creditors from judging the reliability of accounting information. It was suggested that a policy alternative for the K-Unicorn Project, which enhances reliability were presented by reflecting the evaluation of earnings quality through discretionary accruals.

• Korean Journal of Construction Engineering and Management
• /
• v.19 no.1
• /
• pp.65-73
• /
• 2018

Since unfair trading practices between participants in construction project are common, the government has enforced several policies and systems to prevent or minimize the unfair trading practices in construction industry. However, not much attention has been paid to figure out which policies or acts are working or not. This paper analyzed the effectiveness of the policies and acts which are being implemented to prevent unfair trading practices and provided several suggestions to improve the performance of those acts. Survey was conducted to industry experts to collect data regarding their perceptions on those policies and acts. Then the effectiveness of the policies and acts were analyzed in terms of their importance and performance through IPA (Importance-Performance Analysis) based on the survey result. It was found through IPA that execution related acts such as investigation, exposure, and punishment for unfair trading practice have shown low effectiveness in entire construction process and dispute arbitration and mediation related center operated by authority showed low performance too. To improve the effectiveness of those acts, dispute arbitration system improvement, investigation & reporting system consolidation and enhancement practical binding force of punishment and penalty were suggested. Most of all, rules and culture for fair trading should become more established in construction industry by preventing conflict among participants through active communication.

• The Journal of Bigdata
• /
• v.5 no.2
• /
• pp.53-67
• /
• 2020

This study employees a supervised learning prediction model to detect nonconformity in advance of processed food manufacturing and processing businesses. The study was conducted according to the standard procedure of machine learning, such as definition of objective function, data preprocessing and feature engineering and model selection and evaluation. The dependent variable was set as the number of supervised inspection detections over the past five years from 2014 to 2018, and the objective function was to maximize the probability of detecting the nonconforming companies. The data was preprocessed by reflecting not only basic attributes such as revenues, operating duration, number of employees, but also the inspections track records and extraneous climate data. After applying the feature variable extraction method, the machine learning algorithm was applied to the data by deriving the company's risk, item risk, environmental risk, and past violation history as feature variables that affect the determination of nonconformity. The f1-score of the decision tree, one of ensemble models, was much higher than those of other models. Based on the results of this study, it is expected that the official food control for food safety management will be enhanced and geared into the data-evidence based management as well as scientific administrative system.

• The Korean Society of Law and Medicine
• /
• v.22 no.4
• /
• pp.59-88
• /
• 2021

The National Assembly of the Republic of Korea recently enacted laws to fortify sanctions about data manipulation of pharmaceutical companies. The medicine approval system is the result of legislative efforts to prevent accidents that caused damages to patients' life and health. The medicine approval system is based on the trust that the data submitted by pharmaceutical companies is not manipulated. The Supreme Court of Korea clarified that strict standard shoud be required to secure the medicine safety in Supreme Court Decision 2008Du8628 decided November 13, 2008. We agree. This paper suggest legislation to weaken the economic incentives for pharmaceutical companies to choose data manipulation by minimizing the expected profit. In addition to revoking the marketing authorization of the medicine, the 'unfair' profits the pharmaceutical company has earned must be recovered. In addition, in order to increase the possibility to discover data manipulation, it is necessary to strengthen the review capacity and to activate the whistle-blowing.