Constructing ‘ordinary people’ : regulation, production, and representation of ordinary people in the all-star reality game show

Huang, Qiqi Graduate School, Korea University 2022 국내석사

정치적 규제는 중국 텔레비전 산업을 형성하는 데 중요한 역할을 한다. 시진핑 국가주석이 이끄는 정부가 상부구조에 역점을 두면서 미디어 규제기관인 광전총국은 각 분야에 대한 감독과 지시를 강화했다. 특히 올스타 리얼리티쇼가 넘쳐난 텔레비전 시장을 바로잡기 위해서 광전총국은 2015 년과 2017 년에 일반인을 리얼리티쇼에 출연시키는 규정을 발표했다. 이러한 배경에서 본 연구는 중국 대표적인 올스타 리얼리티쇼 <달려라 형제>와 개칭된 <달려라>를 중심으로 일반인 출연자의 재현에 관한 변화를 살펴봤다. 그리고 문화흐름 (the circuit of culture) 이론을 기반으로 규제, 제작, 재현이 어떻게 서로 영향을 미치는지를 탐구했다. 내용분석과 텍스트분석을 통해 일반인 출연자의 확대된 비중, 모범적 정체성, 심층적인 참여 등의 변화가 극명하게 나타난 것으로 밝혀졌고 이는 관전총국의 요구사항을 반영했다는 점도 확인했다. 또한 심층 인터뷰를 통해 각 당사자 간의 투쟁, 협상 및 타협의 과정도 밝혀졌으며, 이는 중국의 상황에서 텔레비전 제작의 복잡한 권력 관계를 보여주었다. Regulations play a vital role in shaping the Chinese television industry, and the media regulator has stepped up supervision on different sectors as the government headed by Xi Jinping has put more weight on the transformation of superstructure. Targeting the television market flooded with all-star reality shows, the regulator issued two regulations specifically requiring the integration of ordinary people into those programs. Under such a context, this study attempted to identify the representational changes of ordinary people in Hurry Up! Brother and its re-titled series Keep Running, the highest-rated reality game show in China, and to explore how regulation, production, and representation entangled with each other in the meaning-making process by adopting the circuit of culture theory as the analytical framework. The content analysis and textual analysis revealed that changes drastically appeared in ordinary people’s larger proportion, professionalized identities, and increased participation, which strongly echoed the regulator’s requirements. However, in-depth interviews also uncovered the intricate process of struggles, negotiations, and compromises among different parties involved, illustrating the complicated power relationships in television production within the Chinese context.

(The) Role of ordinary people in conflict resolution

박연희 경희대학교 평화복지대학원 2017 국내석사

Today's conflicts assume diverse, irregular and complex aspects as seen from terrorism, civil wars, refugees, climate change, corruption of both autocratic and democratically-elected governments, polarization and reemergence of extremism. Whereas institutions and top leaderships are not sufficiently responding to those conflicts with significant achievement, ordinary people occasionally make a great influence to change the conflict situation and contribute to resolve the impasse by campaigning, volunteering, protesting and such. Against this backdrop, this study is an inquiry into the role of ordinary people in alleviating conflict and building sustainable peace. Analytical frameworks for case studies are constructed on the basis of in-depth understanding of ordinary people through the literature review of Silone, Walzer, and Bermeo. The case studies of two villages in Buyeo during the Korean War (Korea) and of Charleston city responding to the tragic church massacre (United States) prove that ordinary people have potential qualities to alleviate conflict in their communities and build foundation for sustainable peace. The study reveals three major findings related to the peacebuilding practices of ordinary people. Firstly, the "experiences" of sharing plight and sharing love or generosity generated strong motivation among people who first offered reconciliation. Secondly, place and religion are significant elements in shaping culturally different practices. Thirdly, ordinary people tried to challenge and transform the underlying unjust social systems beyond the reconciliation between the groups. In conclusion, the study highlights the importance of organic relations between elite and ordinary people and between institution and ordinary people and suggests that the role of ordinary people be understood and exercised in close relation to the elites and institutions.

사회복지사와 일반인의 농 문화 인식 정도에 관한 연구

최정원 순천대학교 사회문화예술대학원 2012 국내석사

본 연구는 사회복지사와 일반인의 농 문화 인식 정도에 관한 연구를 목적으로 하며, 이를 위해 족내혼(농인끼리 결혼), 조직적 네트워크, 사회문화적 관점, 수화, 집단 정체감 및 이 5개를 합친 전체 농 문화의 총 6개 측정항목을 설정하였다. 연구대상은 전남지역 농인관련단체와 복지관 및 시설 등에 근무하는 사회복지사 166명과 일반인 275명이였다. 빈도 및 백분율, 평균, 표준편차, t-test, 교차분석 및 분산분석을 사용하여 분석하였으며 그 결과는 아래와 같다. 첫째, 성별에 따른 항목별 인식의 차이는 없었고, 연령에 따라서 전체 농 문화와 하위 항목에서 인식의 차이를 보였고, 가족이나 친척 중 농인 유무에 따라 농 문화 인식이 차이를 보였으며, 가족이나 친척 중 농인이 있을 경우의 농 문화 인식이 없는 경우보다 더 높았다. 둘째, 사회복지사의 농 문화 인식 특성은 근무경력에 따른 차이가 없었으나, 근무기관에 따라, 청각장애특성 교육 받았는지에 따라, 농인에게 서비스를 제공했는지 여부에 따라 차이를 보였다. 특히 근무기관 중에서는 농인관련단체에 근무하는 사회복지사의 농 문화 인식이 모든 측정 항목에서 가장 높았으며, 농인에게 서비스를 자주 제공한 사회복지사 집단이 농 문화에 대한 인식이 높았다. 셋째, 사회복지사와 일반인 및 수화통역사 집단 사이에 농 문화 인식의 차이가 있었으며, 수화통역사의 농 문화 인식이 가장 높았다. 넷째, 수화학습자의 농 문화 인식은 농인과 만남을 통해 수화를 자연스럽게 습득할 경우에 높았으며, 또한 수화실력, 자격증 유무, 수화사용 경력 및 수화학습 유무에 따라 차이가 나타났다. 인간이 언어를 소유하며 향유함과 동시에 문화가 생성되었다. 수화를 언어로 사용하는 농인 역시 농 문화를 형성하며 장애인이 아닌 소수집단으로서 농 정체성을 확립하고 있다. 따라서 우리는 농 문화를 긍정적으로 인식하고, 이를 위해 농 문화의 대표적 산물인 수화를 보급 할 필요가 있다. 사회복지사는 농 문화의 이해를 통해 농인에게 다양한 사회복지적 서비스를 제공하여 농인의 자존감을 회복하게 함으로써, 농인이 병리적 관점의 청각장애인이 아닌 다문화 환경 속에서 함께 살아가는 동반자적 소수집단임을 알릴 필요가 있다. The purpose of this study was to examine the recognition of deaf culture between the social worker and the ordinary people. A total of 6 measurement items composed of endogamy(marraige between deaf people), organizational network, sociocultural view point, sign language, group identity crisis and the overall deaf culture combining the above 5 items were established to study the recognition of deaf culture by social workers and ordinary people in this research. The subjects were 166 social workers and 275 ordinary people who worked for deaf people-related organizations, welfare centers and welfare facilities in Jeonnam region. The data were analyzed by using the frequency, percentage, average, standard deviation, t-test, chi-square analysis, and analysis of variance. The results were as follows: First, no effect of sex on the recognition was observed but age affected the recognition of the overall deaf culture. The older people's recognition of deaf culture was higher than those of the younger's. The recognition was higher in case they had deaf people in their families than in case they did not. Second, the recognition characteristics of social workers was not affected by the experience period but affected by the kind of organizations they worked for, whether they had received special education for hearing impairment, and whether they had provided any services to deaf people. Especially, the recognition of deaf culture by social workers who worked for deaf-person-related organizations was the highest in all the question items out of the organizations they worked for, and the recognition by the social worker group that provided frequent services to deaf people was high. Third, the recognition of deaf culture was different among groups of social workers, ordinary people, and sign language interpreters and the interpreters showed the highest recognition of deaf culture. Fourth, as far as those who experienced sign language are concerned, the recognition of deaf culture was high for the case that they learned the sign language naturally through direct contact, and it depended on the level of sign language, whether they had the license, and whether they experienced and learned sign language. The moment human beings had and enjoyed their language they created culture. Deaf people who use sign language as their own language also form deaf culture, establishing an identity of a minority group, not the handicapped. Thus, it is necessary to recognize the deaf culture positively and publicize sign language, the representative product of deaf culture. Social workers need to inform deaf people that they are a minority group of companions living together in multi-cultural environment, not hearing-impaired persons in pathological view point by providing a variety of social welfare services to them though understanding of deaf culture and having them recover their self-esteem.

A Platform for Diagnosing the Prevalence of Multiple Diseases in Ordinary People Using an Explainable Machine Learning

ZHENG, HUILIN 충북대학교 2023 국내박사

연구 동기 및 목표: 전 세계적으로 일반인의 다중질환 위험이 증가하고 있으며, 사망자 수가 크게 증가하고 있어, 증상이 없더라도 정확하고 효율적인 질병 예측이 중요하고 필요하다. 의료 분야에서 다양한 질병을 예측하기 위해 기계학습과 딥 러닝 기술을 기반으로 많은 진단 모델이 개발되었으며 기존의 통계 기반 방법보다 더 높은 성능을 달성했다. 그러나 설계된 모델의 구조가 불분명하고 일반인의 다중질환을 쉽게 예측할 수 있는 플랫폼이 아직 부족하다. 따라서 본 논문에서는 설명 가능한 기계학습 기법을 이용하여 일반인을 대상으로 한국국민건강영양조사 (KNHANES) 데이터셋에서 고혈압, 당뇨병, 이상지질혈증, 심근경색, 협심증 등 다중질환 유병률을 진단할 수 있는 플랫폼을 제안하고자 한다. 연구방법 및 자료: 논문의 세부 연구 방법은 다음과 같이 요약할 수 있다. 먼저, 본 논문에서 2007년부터 2020년까지 14년간의 KNHANES 데이터셋을 병합하고, 상관관계가 없는 변수 추출, 결측 값 처리, 데이터 정규화 등 다양한 방법을 사용하여 데이터셋을 전처리한다. 그 다음에 전처리된 데이터셋을 고혈압, 당뇨병, 이상지질혈증, 심근경색, 협심증 등 각 질환별로 5가지 하위 집합으로 나누고 무작위 언더샘플링 기법을 적용하여 질병이 없는 일반인의 개체 수를 줄인다. 이후 각 부분 집합은 80%의 훈련 데이터와 20%의 테스트 데이터로 분할된다. 둘째, 본 논문에서 각 하위 집합에 여러 개의 단일 기계학습 알고리즘과 앙상블 분류기를 적용하여 각 질병에 가장 적합한 모델을 찾는다. 단일 알고리즘은 Naïve Bayes, 로지스틱 회귀, 서포트 벡터 머신, 의사결정 트리, 심층 신경망 등을 포함하고 앙상블 알고리즘은 random forest, AdaBoost, XGBoost, LightGBM, CatBoost 과 스태킹 방법으로 구성된다. 한편, 모델 학습 과정에서 개발된 모든 모델의 과적합을 방지하기 위해 5배 교차 검증이 사용되며, 모든 ML 기반 모델의 하이퍼 파라미터는 베이지안 최적화를 통해 조정한다. 셋째, 테스트 데이터를 사용하여 학습된 모델의 성능을 평가하고 정확도, 정밀도, 리콜, F1 점수, 기하학적 평균, 수신기 작동 특성 곡선 아래 면적(AUC), Matthews 상관 계수(MCC) 및 해밍 손실 등의 성능 메트릭을 사용하여 실험 결과를 비교한다. 넷째, 기계학습 설명 기법(DALEX)을 사용하여 각 질병의 주요 위험 요인을 파악하고, 최적의 모델을 설명하기 위해 국소 해석 가능한 모델 독립적 설명 기법(LIME)을 적용한다. 마지막으로, 각 질병의 주요 위험 요인을 기반으로 일반인의 다중질환 위험을 예측하기 위한 웹 기반 다중질환 진단 플랫폼을 개발한다. 결과 및 토의: 종합적인 실험 결과는 각 질병에 대한 최적의 예측 모델의 성능이 여러 질병의 예측에서 다른 기계학습 및 딥 러닝 기반 모델의 성능을 능가한다는 것으로 나타났다. 이상지질혈증 진단에서는 XGBoost 모델이 0.8715의 정확도로 가장 우수한 결과를 보였고, 심근경색과 고혈압 진단에서는 스태킹 모델이 각각 0.9167과 0.9059의 정확도로, 협심증과 당뇨병 진단에서는 CatBoost 모델이 각각 0.8908과 0.961의 정확도로 가장 우수한 성능을 나타냈다. 결론: 결과적으로 제안된 진단 플랫폼은 일반인의 다중질환 유병률을 정확하게 예측할 수 있으며, 각 질병에 대한 최적의 예측 모델에 대한 설명은 블랙박스 모델에 대한 이해도를 높일 수 있다. Research Motivation and Objectives: The risk of multiple diseases in ordinary people is increasing and has significantly increased the number of deaths worldwide, so more accurate and efficient predictions of diseases in ordinary people are important and necessary even if they do not exhibit any symptoms. Many diagnosis models based on machine learning and deep learning techniques have been developed to predict different diseases in the medical field and have achieved higher performance than conventional statistics-based approaches. However, the structures of those prediction models are unclear, and it still lacks a platform that can be easily used to predict multiple diseases in ordinary people. Therefore, this paper proposes a platform for diagnosing the prevalence of multiple diseases such as hypertension, diabetes, dyslipidemia, myocardial infarction (MI), and angina in ordinary people using an explainable machine learning technique in the Korea National Health and Nutrition Examination Survey (KNHANES) dataset. Research Methods and Material: The detailed research methods of the paper can be summarized as follows. First, this paper merges a 14-year KNHANES dataset from 2007 to 2020 and preprocesses the dataset using different methods including uncorrelated variable extraction, missing value processing, and data normalization. Then the preprocessed dataset is divided into five subsets with each disease: hypertension, diabetes, dyslipidemia, MI, and angina, and the ordinary people without any disease is randomly sampled to reduce the number of individuals based on the number of datasets for each disease. Thereafter, each subset is split into training data of 80% and test data of 20%. Second, this paper applies multiple single machine learning algorithms and ensemble classifiers to each subset to find the best model for each disease, where the single algorithms include Naïve Bayes, logistic regression, support vector machine, decision tree, and deep neural network, while the ensemble algorithms contain random forest, adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), CatBoost, and stacking methods. Meanwhile, 5-fold cross-validation is employed to prevent overfitting in all developed models throughout the model training processes, and the hyperparameters of all ML-based models are adjusted via Bayesian optimization. Third, the performance of the trained models is evaluated using test data and the experimental results are compared using the following performance metrics: accuracy, precision, recall, F1 score, geometric mean, the area under the receiver operating characteristic curve (AUC), Matthews correlation coefficient (MCC), and hamming loss. Fourth, the descriptive machine learning explanations technique (DALEX) is used to identify the key risk factors for each disease and the local interpretable model-agnostic explanations (LIME) is applied to explain the best models. Lastly, a Web-based multiple diseases platform is developed to diagnose the risk of multiple diseases for the ordinary people based on the key risk factors for each disease. Results and Discussion: The comprehensive experimental results indicated that the performance of optimal prediction models for each disease outperformed those of other machine learning and deep learning-based models in the prediction of multiple diseases. The XGBoost model provided the greatest results for diagnosing dyslipidemia with an accuracy of 0.8715; the stacking models achieved the best performance for MI and hypertension diagnosis with the accuracy of 0.9167 and 0.9059, respectively; and the CatBoost models performed the best performance to diagnose angina and diabetes with the accuracy of 0.8908 and 0.961, respectively. Conclusion: Consequently, the findings of this paper were that the proposed platform could diagnose the probability of the prevalence of multiple diseases in ordinary people accurately and the explanation of the optimal prediction models for each disease should also improve the understanding of the black-box prediction models.

역사학습에서 ‘보통사람’에 대한 인식

이현아 서울대학교 대학원 2008 국내석사

A Systems Approach to Elucidate Personalized Mechanistic Complexities of Antibody-Fc Receptor Activation Post-Vaccination

Lemke, Melissa Marie ProQuest Dissertations & Theses University of Mich 2022 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

One of the most significant challenges to current human healthcare is the emergence of antigenically variable viruses that evade traditional vaccination approaches. Human immunodeficiency virus (HIV) is one such virus that emerged over 30 years ago and still has no effective vaccine. Like many other antigenically variable viruses, after infection, HIV quickly mutates to evade broadly neutralizing antibodies that bind tightly to key sites to prevent infection. Over 250 clinical trials have been performed to date to develop an effective HIV vaccine, with only one providing moderate protection; the RV144 Thai trial, estimated to be 31% effective but has not been replicated in other populations. Rather than broadly neutralizing antibodies, the trial identified IgG antibodies with the capacity to induce Fc effector functions as a correlate of protection. These functions are triggered by less specific antibodies that bind HIV antigen and Fc receptors on the surface of innate immune cells to form immune complexes to activate protective cellular functions. Understanding how to increase the formation of IgG-FcR complexes may improve vaccine efficacy, but variation in IgG and FcR features across individuals suggests that protective mechanisms need to be understood on a personalized basis. There are multiple subclasses of protective IgGs, each having different concentrations and affinities to FcRs in different individuals. Genetics can also play a role, with FcR polymorphisms changing FcR binding affinity and IgG1 allotypes changing IgG subclass concentrations. Mechanistic ordinary differential equation (ODE) modeling of this system offers the opportunity to account for these factors on a personalized basis and deconvolve which are most influential and determine how to improve protection universally. We developed an ODE model of IgG-FcγRIIIa immune complex formation to elucidate how personalized variability in IgG subclass concentration and genetic factors may contribute to complex formation after vaccination. We validated the model with RV144 plasma samples and used it to discover new mechanisms that underpin complex formation. This enabled the identification of genetic and post-translational features that influenced complex formation and suggested the best interventions on a personalized basis. For example, although IgG3 was associated with protection in RV144 and has the highest affinity to FcγRIIIas, the model suggested that IgG1 may play a more essential role, though it also may be highly variable; due to high IgG1 concentration variability across individuals. The model identified RV144 vaccinees who were predicted to be sensitive, insensitive, or negatively affected by increases in HIV-specific IgG1, which was validated experimentally with the addition of HIV-specific IgG1 monoclonal antibodies to vaccine samples. The model also gave important insights into how to maximize IgG-FcγRIIIa complex formation in different genetic backgrounds. We found that individuals with certain IgG1 allotypes were predicted to be more responsive to vaccine adjuvant strategies that increase antibody affinity (e.g., glycosylation modifications) compared to other allotypes, which were predicted to be more responsive to vaccine boosting regimens that increase IgG1 antibody concentration. Finally, simulations in mixed-allotype populations suggest that the benefit of boosting IgG1 concentration versus IgG1 affinity may depend upon the frequency of a specific IgG1 allotype (G1m-1,3) in the population. Overall we believe that this approach represents a valuable tool that will help understand the role of personalized immune mechanisms in response to vaccination and address challenges related to under-represented genetic populations in vaccine trials.

(A) study on the hair analysis of a nutritious mineral and a toxic heavy metal for the elderly with senile dementia for the creation of clean community

안붕혁 Graduate School, Yonsei University 2015 국내박사

중국 소설을 통해 본 '개인'에 대한 인식의 변화 : '四大奇書'부터 『紅樓夢』까지의 변화를 중심으로 : from sida qishu to Hongloumeng

최형섭 서울대학교 대학원 2007 국내박사