호 지역성을 고려한 분산 계층 SCP 구조에서의 PCS 이동 관리 기법

양창용 연세대학교 1996 국내석사

유해 중금속의 생물분류군별 생태독성평가

양창용 건국대학교 대학원 2009 국내석사

중금속은 환경 중 자연적으로 존재하는 물질이며 산업이 발달함에 따라 다량의 중금속이 생태계로 유입되어 급․만성 독성영향을 초래할 수 있다. 한편 인체의 경우 중금속에 오염된 동․식물을 섭취하거나 기타 여러 노출경로에 의해 노출될 경우 미량이라도 체내에 축적되면 잘 배설되지 않고, 단백질에 쌓여 장기간에 걸쳐 부작용을 가져올 수 있어 매우 위험하다. 이러한 중금속이 생태계로 유입되어 생물체에 미치는 독성영향을 알아보고자, 수서 및 토양매체에서 서식하는 생물종인 Oryzias latipes (Fish), Caenorhabditis elegans (Nematode), Fridericia peregrinabunda (Enchytraeidae) 그리고 Pseudokirchneriella subcapitata (Green algae)를 대상으로 하여 독성영향에 대해 평가하였다. 대상 중금속은 안티몬, 크롬, 카드뮴, 아연, 니켈, 구리, 수은, 비소(3가 및 5가), 납, 나노실리카 등 11종류의 중금속을 대상으로 하였다. 독성종말점은 다양한 생물종을 이용하여 얻은 결과를 LC50, EC50, NOEC, LOEC로 산정하였다. 또한 C. elegans을 이용한 복합독성실험에서 복합독성모델 (TU, Toxic Unit)을 적용하였다. 본 연구결과, 어류 검정법에서 O. latipes, larvae를 이용한 독성실험 결과 성어보다 자어나 수정란의 생육단계에서 독성의 민감성이 크게 나타난 것을 알 수 있다. 특히 중금속 중 수은, 구리, 카드뮴, 아연 등 독성이 크게 나타났다. 선충류 검정법에서 C. elegans를 이용한 독성실험 결과 수은, 납, 아연 등 다른 중금속에 비해 독성이 큰 것으로 나타났으며 어류 검정법과 비슷한 양상을 보였다. 복합독성실험에서는 모든 조합(구리, 카드뮴, 납)에서 1TU 이하로 산정됨으로써 독성영향의 상승작용이 일어나는 것을 알 수 있었다. 애지렁이류 검정법에서 F. peregrinabunda를 이용한 독성실험 결과 카드뮴의 경우, 48(72)시간 LC50은 21.02(18.88) mg L-1, 안티몬의 경우, 24(48)시간 LC50은 387(37.21) mg L-1을 산출되었다. 조류의 성장저해시험에서 안티몬은 72시간 EC50 206 mg L-1로 측정 되었으며 N(L)OEC 100(150) mg L-1을 산출되었다. 한편 나노실리카는 수정란에 대한 발달독성을 나타내었으며, 나노실리카가 생물체내부로 침투됨을 확인하였다. 본 연구에서 사용된 생물분류군별 생태독성평가는 중금속에 대한 생태독성을 광범위하게 파악하였으며, 중금속과 생물종의 상관관계의 이해에 도움이 된다고 판단된다. Heavy metal is a naturally occurring material in the environment, and the inflow of a large amount of heavy metal to the ecosystem as the industrial wastewater cause acute and chronic toxic effects. In the present study, ecotoxicity of several heavy metals was investigated using a battery of bioassay. The test species used are Oryzias latipes (Fish), Caenorhabditis elegans (Nematode), Fridericia peregrinabunda (Enchytraeidae), and Pseudokirchneriella subcapitata (Green algae). Test chemicals include antimony, chromium, cadmium, zinc, nickel, copper, mercury, arsenic (3+, 5+), lead, and nano silica. Toxicity endpoints such as LC50, EC50, NOEC, and LOEC were calculated. Mixture toxicity of Cd, Cu, and Pb was assessed by using C. elegans based on Toxic Unit approach. In addition, developmental toxicity and bioaccumulation of nanoscale silica was evaluated by a fish embryo bioassay. Larval assay using O. latipes, was more sensitive than adult fish assay. Mercury was the most toxic to fish assay, and copper and cadmium was also very toxic. Copper and antimony show the developmetal inhibition to the fertilized egg of O. latipes. In the nematode assay using the C. elegans, mercury, lead, zinc and other heavy metals found toxic. In the mixture toxicity test using Cd, Cu, and Pb, TU was calculated as less than 1, indication synergistic effect of three heavy metals. Enchytraeidae assay using F. peregrinabunda was performed. It was found that F. peregrinabunda was a potential bioindicator in soil pollution. Growth of P. subcapitata was inhibited by antimony. In addition, toxicity of nano silica for development of fertilized egg was observed, and, nano silica was infiltrated into the egg membrane. This study will give a better understanding about the toxicity of toxic heavy metals by using a battery of bioassay. It is also useful to understand the correlation between heavy metals and species.

제7차 사회과 교육과정에서의 영상 미디어 활용 교육 방안 연구 : 고등학교 일반사회 교과 영역을 중심으로

양창용 전남대학교 교육대학원 2004 국내석사

본 논문은 제7차 사회과 교육과정에서 영상 미디어를 어떻게 활용하여 교육할 것인가에 대한 방안을 고등학교 일반사회 교과영역 중심으로 연구하였다. 현대 사회는 정보화, 세계화, 개방화의 물결속에서 급격하게 변하고 있다. 제7차 교육과정은 급변하는 사회를 주도할 자율적이고 창의적인 민주 시민 교육을 지향하고 있으며, 사회과는 이러한 사회변동에 대응하는 핵심적인 교과이다. 고등학교 사회과에서는 사회현상을 종합적으로 이해하고, 비판적 사고와 합리적 의사 결정 능력을 발휘하며, 사회 공동 문제 해결에 적극적으로 참여하는 시민의식을 기르는 데 중점을 두고 있다. 지금까지 사회과 수업은 지루한 주입, 암기 위주의 교과라는 비판을 줄곧 받아오고 있다. 사회과는 다양한 사회현상에 관심을 갖는 교과로서 학습자들이 다양한 사회현상을 직접 보고, 듣고, 경험할 수 있는 학습의 장을 제공하는 것이 중요하다. 지구화로 세계의 각 지역학습, 문화학습, 환경학습 등에 말보다 실제의 모습을 담은 영상 미디어가 의미있음은 말할 필요가 없다. 이러한 학습은 시공간 및 비용부담의 한계성을 극복할 수 있어 간접적으로 생동감 있는 경험의 장을 제공해 줄 수 있게 됨으로써 새로운 영상 미디어의 활용이 적실(適實)하게 되었다. 21세기의 사회변화에 잘 적응하고, 더 나아가 변화를 주도할 수 있는 인간을 양성하기 위해서는 학습내용의 구성과 학습방법 및 평가 방법 등에 있어서도 혁신적인 변화가 요구되며, 이러한 학습방법과 학습자료에 중요한 역할을 하는 것이 영상 미디어이다. 교육의 각 분야에서도 패러다임 변화에 따른 학습 매체로서 영상 미디어에 대한 관심이 점차로 높아지고 있다. 따라서 이것을 교육적으로 활용하기 위한 방안을 모색하였다. 시대에 부합하는 매체의 하나로서 영상 미디어를 교육과정에 뿌리내리고 제몫을 다하기 위해 영상 미디어의 교육적 가치는 무엇인지, 교과영역으로서 사회과 교수·학습에 어떠한 영상 미디어 자료를 개발하여 활용할 것인지 등을 알아보았다. 영화를 활용한 사회수업, 시사 애니매이션을 활용한 사회수업, 뉴스를 활용한 사회수업, 시사 다큐멘타리를 활용한 사회수업, 공익광고를 활용한 사회수업 모델을 연구하였다. 이러한 점을 감안하여 본 연구에서는 영상 매체 속에 살아가는 현실의 삶의 연장선상에서 어떠한 영상 미디어를 개발하여 사회과 교육에 효율적으로 활용할 수 있을지를 모색하였다. 이를 위해 영상 미디어의 특성과 교육적 가치를 관련문헌을 통해 검토하고, 영상 미디어를 교육에 활용하기 위해 어떻게 매체를 개발하여, 사회과 교육에 활용할 것인가를 살펴보았다. This paper studied teaching methods of utilizing the visual media in social studies in the 7th curriculum on the focus of general social studies in high school. The modern society has been changing rapidly in the waves of information, globalizat ion and openness to the world. The 7th curriculum aims to educate self-controlling, creative and democratic citizen to lead rapidly-changing society and social studies is the core subject for responding to those social currents. The aim of social studies in high school is to bring up the citizenship with which citizen can generally understand social phenomena, make critical considerations and rational decision and voluntarily participate in solving social problems. So far, class of social studies has been criticized for boredom and memorization-oriented learning. Social studies is the subject which reflects a variety of social changes and it is important to provide students with the educational field where they can see, listen and experience various social phenomena. It is needless to say that authentic material which contains the real world such as the world regions, cultures and environments is much more meaningful than just words or let ters. Accordingly new teaching methods of visual media meet the need to supply students with lively field of experience indirectly because it can overcome the limit of time, space and expenses. A radical change of the learning contents, teaching methods and evaluations is needed in producing human being who can adapt themselves in the 21st century's s social changes and lead the innovations and the visual media play major role in learning met hods and materials. In every field of education, visual media as a learning medium draw increasing at tentions according to changes of paradigm. So in this study various methods of putting visual media into education was investigated. Also the educational value of visual media was studied in the curriculum as a means to catch up with the times. And what kinds of visual media material to develop and to utilize into social studies' learning and teaching was examined. This investigated the class model through movies, current animations, news, current documentary programs. On aspects of this, this research aimed to develop educational visual media and to find out the effective ways to apply visual media into social studies education in a world of mass media. This paper examined the properties of visual media and the educational value through the related articles and surveyed how to develop media and how to apply it into social studies education.

퍼지理論을 이용한 트러스 構造物의 形狀最適化에 관한 硏究

양창용 전북대학교 대학원 1997 국내박사

Optimum design of structure is categorized into member sizing optimization for fixed shape and geometric shape optimization(GSO). However, GSO makes it possible to reduce both an weight of structure and cost. To date, multi-level optimization(MLO) approach has been the primary focus of discussions regarding efficient GSO as the structure has been complexed and enlarged. A variety of functions could be utilized for making optimum design to functionally different structures, while those characteristics are really compensated each other. However, multi-objective optimization is not only consisted of many design variables and constraints, but also it is not possible to form a simple optimization process due to nonlinear constraints and different characteristics. Futhermore, a designer could face various difficulties caused mainly by fuzzy constraint and goal when he develops an exact mathematical model for certain structure. From a standpoint for those fuzzy characteristics, therefore, optimizing constraints and objective should include fuzziness, and it presents necessity to related research. The research project described in this report had an overall objective to evaluate optimization of multi-objective function(MOF) and truss structure including fuzziness within constraint. Also included is MLO approach concerning partition into the first level and the second level optimization approaches for overall process of optimization in order to perform effective shape optimization(SO). The first stage of optimum design was performed using cross-sectional area and fuzziness as the design variable and the objective function, respectively. Considering stress constraints, the fuzziness, GSO was formulated using counteractive volume and displacement as the first objective function and the second objective function, respectively. Optimum solution was also formed using multi-objective fuzzy optimization through feasible direction method. For the second stage of optimum design, Powell method was applied to form optimum shape using configuration or geometric layout variables, volume, and displacement as the design variable, the first objective function, and the second objective function, respectively. Consequently, this shape optimization included two counteractive objective functions as one fuzziness and constraints for 2 bars, 13 bars, and 18 bars as another fuzziness. The results are as follows ; 1. Using MLO approach, this research made it possible to make SO of truss structure consisted of counteractive MOF and constraints as the fuzziness. 2. Decreasing the number of constraints and design variables caused simplification for optimization process by partition of optimization stage. 3. When applied to three kinds of truss structure including the MOF as one fuzziness and the constraints as another fuzziness to be fixed to the optimization shape, the algorithm in this research decreased volume and displacement as much as 1.8 to 13.3% and 1.5 to 7.39%, respectively. Associated with the algorithm was the convergence of iteration within four times. Consequently, this approach would be effective as far as iteration is concerned. 4. The algorithm in this study could be suitable for optimum design for complex structure including MOF and constraints as the fuzziness.

판별분석과 머신러닝 알고리즘을 이용한 중소기업의 부실예측모형에 대한 연구 : 비외감기업을 중심으로

양창용 숭실대학교 대학원 2023 국내박사

본 논문은 그동안 상장기업 등 위주로 연구된 부실예측모형에서 한걸음 더 나아가 재무비율을 이용하여 중소기업의 부실예측모형을 개발하고 중소기업의 부실에 따른 다양한 손실을 줄이기 위한 지표로 활용할 수 있는 정보를 제공하는데 그 목적이 있다. 중소기업 종합지원기관인 신용보증기금에서 축적하고 있는 신뢰도 높은 중소기업의 재무제표를 표본으로 하여 총 63,874개를 분석에 사용하였으며 이중 제조업은 31,623개로 49.5%, 도소매업은 20,443개로 32.0%, 건설업 등 기타 업종이 11,808개로 18.5%를 차지하였다. 이에 본 논문에서는 전체업종 및 비중이 큰 제조업과 도소매업을 분석대상으로 하였다. 기업의 부실 여부를 예측할 수 있는 독립변수로는 신용보증기금에서 중소기업에 대한 보증지원 심사시 활용하고 있는 21개의 재무비율을 사용하여 판별분석을 통해 부실예측모형을 도출하였다. 전체 표본을 대상으로 한 분석에서는 21개의 재무비율 중 ‘자기자본비율’ 등 판별력이 높은 15개의 재무비율을 이용하여 판별함수를 산출하였으며 판별함수가 표본기업들의 분류를 얼마나 잘 예측하는가를 보여주는 hit ratio는 95.7%였다. 전체 표본 중 32.0%를 차지하고 있는 도소매업을 영위하는 기업의 재무제표에 대한 분석을 별도로 실시하여 21개 재무비율 중 ‘금융비용부담률’ 등 13개의 판별력이 높은 변수를 선정하였다. 13개의 재무비율을 이용하여 산출한 판별함수의 hit ratio는 95.2%였다. 또한 전체 표본 중 약 절반가량인 49.5%를 차지하고 있는 제조업을 영위하는 기업의 재무제표에 대한 분석도 별도로 실시하여 21개의 재무비율 중 ‘총자본세전계속사업이익률’ 등 10개의 판별력이 높은 변수를 선정하였다. 10개의 재무비율을 이용하여 산출한 판별함수의 hit ratio는 91.3%였다. 머신러닝 기법을 활용하여 추가적인 분석을 실시하였다. 판별분석은 모형 자체와 독립변수의 통계적 유의성을 검증할 수 있다는 장점은 있으나 분류정확도는 상대적으로 머신러닝 알고리즘 대비 낮으므로, 일반적으로 더 높은 분류정확도를 나타내면서도 각 독립변수별 중요도를 계산할 수 있는 머신러닝 기법들인 XGBoost, LightGBM, 랜덤포레스트(Random Forest)를 활용하여 정상기업과 부실기업인 종속변수를 예측하고 재무비율인 독립변수별 중요도를 계산하였다. XGBoost, LightGBM, 랜덤포레스트 모델의 정확도, 정밀도, 재현율, F1점수 모두 0.97∼0.98의 값을 보여 판별분석과 비교했을 때 아주 높은 수준의 예측력을 보여주었다. AUC 또한 세 모델 모두에서 0.89∼0.90으로 매우 높은 수준이었다. 독립변수별 중요도는 ‘총자본세전계속사업이익률’, ‘금융비용부담률’, ‘비유동장기적합율’, ‘부채비율’, ‘자기자본비율’ 순으로 중요도가 높아 정상기업과 부실기업을 구분하는데 영향을 미치는 것으로 분석되었다. 이는 중요도의 순서는 다르지만 판별분석의 분석력이 높은 상위 5개의 독립변수와 동일함을 알 수 있었다. 머신러닝 기법으로 중소기업의 재무비율을 이용하여 정상기업과 부실기업을 분석하여 기존 통계방법론을 보완하고 정상기업과 부실기업으로 분류하는 예측정확도를 높일 수 있었을 뿐만 아니라 기존 샘플링 자료에 대한 새로운 통찰을 획득하였다는 것이 선행연구와의 차별점이라 할 수 있다. The purpose of this paper is to take a step further from the default prediction model studied mainly for listed companies, etc., to develop a default prediction model for SMEs using financial ratios and to provide information that can be used as an indicator to reduce various losses caused by default. A total of 63,874 reliable financial statements accumulated by the Korea Credit Guarantee Fund, a comprehensive support organization for SMEs, were used for the analysis. Among them, manufacturing accounted for 49.5% with 31,623, wholesale and retail accounted for 32.0% with 20,443 and other industries such as construction with 11,808 accounting for 18.5%. The 21 financial ratios used by the Korea Credit Guarantee Fund for screening guarantee support for SMEs were used as independent variables to derive a default prediction model through discriminant analysis. In the analysis of all samples, the discriminant function was calculated using 15 highly discriminating financial ratios, such as ‘capital adequacy ratio’ and the hit ratio, which shows how well the discriminant function predicts the classification of sample companies, was 95.7%. A separate analysis of the financial statements of companies operating in wholesale and retail, which account for 32.0% of the total sample, was conducted to select 13 highly discriminating variables, including the ‘financial cost burden ratio’, among the 21 financial ratios. The hit ratio of the discriminant function calculated using 13 financial ratios was 95.2%. In addition, a separate analysis of the financial statements of manufacturing companies, which account for about half of the total sample, 49.5%, was conducted to select 10 highly discriminating variables such as the ‘ratio of operating profit to total capitals’. The hit ratio of the discriminant function calculated using 10 financial ratios was 91.3%. Additional analysis was conducted using machine learning techniques. Discriminant analysis has the advantage of verifying the statistical significance of the model itself and independent variables, but the classification accuracy is relatively lower than that of machine learning algorithms. Therefore, XGBoost, LightGBM and Random Forest, which can calculate the importance of each independent variable while showing higher classification accuracy, were used to predict normal and default companies and calculate the importance of each independent variable, which is the financial ratio. The accuracy, precision, recall and F1 score of the XGBoost, LightGBM, and Random Forest models all showed values of 0.97 to 0.98, showing a very high level of predictive power compared to discriminant analysis. AUC was also very high in all three models, ranging from 0.89 to 0.90. The importance of each independent variable was analyzed to be high in the order of ‘ratio of operating profit to total capitals’, ‘financial cost burden ratio’, ‘non-current long-term conformity rate’, ‘debt to equity rate’ and ‘capital adequacy ratio’, which affect the distinction between normal and default companies. Although the order of importance was different, it could be found that the top five independent variables with high analysis power of discriminant analysis were same. By analyzing normal and default companies using the financial ratio of SMEs through machine learning techniques, it was possible to supplement the existing statistical methodology and improve the prediction accuracy of classifying them into normal and default companies, as well as obtaining new insights on existing sampling data. These are the difference from previous studies.

제주방문 관광객의 펜션이용 특성에 관한 연구

양창용 제주대학교 2004 국내석사

離散型 斷面을 考慮한 트러스構造物의 最適設計에 관한 硏究

양창용 全北大學校 1993 국내석사

A discrete formulation of the optimum design structures is developed and methods of solution are presented. The design problem is the allocation of member sizes from catalog of commercially available sections within the working stress constraints, the buckling stress constraint and the displacement constraint. Multiple sets of static service loads are considered in the formulation. A matrix set of the constraints of working stress, buckling stress and disp1acement is derived by the stiffness method, following the networktopological approach and incorporating the working stress interaction conditions. Constraints for optimization of truss structures is transformed into the constraint set of a successive binary programming problem. The enumeration method developed by Balas is considered to he one of the most efficient algorithms for binary programming and therefore is chosen as the basic algorithm for this study. An improved algorithm which utilizes the multiple choice constraints characteristic of discrete structural design is developed. A computer program is developed for the implementation of the above formulations and algorithm. Example application include two bar truss, seven bar truss and ten bar truss,

고차원 색인 구조를 위한 벌크 삽입 알고리즘의 설계

양창용 충북대학교 산업대학원 2001 국내석사

고등학교 영어교과서에 나타난 동사구 연어의 코퍼스 기반 연구

고혜정 제주대학교 교육대학원 2012 국내석사

Topicality in Korean - a discourse-functional approach : a case of -Nun, -Nuntey, -Isscanha

양창용 University of Florida 2001 해외박사

The present study examines some topic expressions and their dynamic functions in Korean, which are realized by such various linguistic coding devices as -nun, -nuntey, and -isscanha. Previously, topic studies in Korean have concerned themselves with the identification of a topical entity drawing on the topic marker -nun. Aware of the limitations of previous analyses of the topic structure, we take up the issue from a different point of view: a discourse/functional perspective. Our underlying assumption is that the choice of a topical form representing an entity or event is interactively determined by dynamic discourse contexts. This study has two specific assumptions within the discourse/functional framework. One is that various topical expressions such as a -nun-marked NP and ZA are used to convey different discourse/pragmatic implications. The other is that the topic as a dynamical discourse notion is introduced, developed and continuously organized within situational contexts. For the first, we adopt Chu’s markedness model consisting of a continuum of two typical topic structures: marked vs. unmarked. Based on natural discourse data, we examine the typical topic realizations in Korean: the -nun topic and zero anaphora. It is argued that the -nun topic, as a marked one, is linguistically prominent, viz. high in informative value. In addition, the concept of marked topic is useful in the sense that it can offer some insight into how focus reading sometimes co-exists with topic reading. On the other hand, ZA, as an unmarked topic, is low in informative value and serves a unique function of linking clauses, a function which marked topics lack. For the second assumption, we investigate two linguistic devices that can be readily regarded as what is called discourse marker: -nuntey and -isscanha in terms of topic-organizing processes. The claim made is that topic should be viewed from the how-perspective in discourse organizing processes. In other words, participants use various linguistic devices to signal how their contributions are related to prior and subsequent segments of discourse. It is shown that two discourse markers -nuntey and -isscanha come into play in helping various actual topic interpretations. These functions are derived from their fundamental properties, viz. -nuntey as a non-finality marker and -isscanha as an attention-calling device. More specifically, -nuntey, signaling more information is yet to come, has to do with such actual interpretations as topic-introduction, topic-resumption, topic-continuation, and topic negotiation. -Isscanha, attracting attention to what immediately follows the utterance it occurs in, helps the speaker facilitate dynamic topic developments. Among its actual interpretations are topic introduction, topic establishment, topic seeking, topic shift and listing of topics.