http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 저자 : 강동주 (검색결과 173 건)
- 무료
- 기관 내 무료
- 유료
-
-
-
강동주 한국경찰복지연구학회 2014 경찰복지연구 Vol.2 No.2
유럽, 북미, 아시아 등 많은 국가들은 이미 오래전부터 경찰수사에 경찰수사목 적견을 활발하게 사용하여 다양한 분야에서 성과를 거두고 있으며 지속적인 연구 가 이루어지고 있다. 이에 우리 경찰청은 2012년 11월부터 공식적으로 경찰수사 목적견을 9개 지방경찰청에서 도입 운영 중에 있고 경찰수사목적견 도입을 희망 하는 지방경찰청과 경찰서가 점차적으로 증가하고 있는 추세이다. 하지만 경찰수 사목적견 도입 초기인 한국 경찰은 경찰수사목적견에 대한 이해 및 인식의 부족, 연구의 부재, 경찰수사목적견 운영과 관련한 문제점 등 개선해야 할 부분이 많이 산재해 있어 이에 대한 해결방안이 시급한 실정이다. 이러한 연유로 본 연구가 수 행되었으며 본 연구의 목적은 경찰수사목적견 운영상의 전반적인 실태를 진단하 고 문제점을 분석하여 경찰수사목적견의 활성화 방안을 제시하는 데에 있다. 본 연구의 의의는 국내 경찰수사목적견을 도입하면서 본 연구자가 담당 핸들러 로서 현장속에서 부딪히는 생동감 있는 상황을 주관적으로 경험하였을 뿐만 아니 라 동료 핸들러와 형사 및 타기관 전문가들의 경험과 시각으로 보다 객관화된 활 성화 방안을 탐구하였다는데 있다. 또한 경찰수사목적견은 본 연구를 시작으로 지 속적인 연구가 수행되어지고 발전하기를 기대한다. The purpose of the current study is to provide useful information for enhancing the effectiveness of police dogs related to criminal investigation. Police dogs have long been used in Europe, North America, and Asia. Since the police dog used for the purpose of criminal investigation was officially introduced in Korea in November, 2012, the demand for police dog has steadily increased. However, in spite of several barriers to development of police dogs, research on police dogs has not been conducted. To investigate the current situation of Korean police dogs in criminal investigation, official data of related agencies, government reports and internet information are analyzed. Overseas examples are also reviewed to elicit implications of any related experiences. In addition, a qualitative interview method is employed. Twenty one experts including police dog expert, other agencies' investigation dog experts, police officers, professors, criminal investigation experts and veterinarians participated in the interview. Based on previous research, police dog for investigation, police dog handlers and working environment are used for research topics. Based on research findings, 8 policy implications are presented. First, a police dog task force should be made to enhance the expertise. Second, a proper selection process needs to be developed to recruit qualified handlers. Third, a purchasing channel for police dog needs to be diversified to gain better police dogs. Fourth, the perception of the police towards police dogs should be improved. Fifth, the police dog training program should be developed. Sixth, facilities such as a training camp for the management of police dogs needs to be improved. Seventh, developing legal system for the activities of police dog and taking insurance are needed. Lastly, welfare system for the police dog need to be developed. As the very first academic research on police dogs for the purpose of criminal investigation, the current study is expected to provide preliminary information to enhance the effectiveness of police dogs.
-
흡수된 상악구치부에서 임프란트 이식방법에 따른 응력분포 비교
강동주,임장섭,정창모,전영찬,임오강 대한치과이식임플란트학회 2001 The Korean Academy of Implant Dentistry Vol.20 No.1
The purpose of this study was to compare the stress distribution of the different treatment strategies for reconstruction of the resorbed posterior maxilla with implants. A three dimensional finite element analysis method was used to investigate these strategies. Six different finite element models, composed of cortical and trabecular bone involving a single implant, were constructed. Model 1 was monocortication by crestal cortical bone engagement only using a 3.75mm wide and 7mm long implant. Model 2 was bicortication of crestal cortical bone and sinus floor cortical bone using a 3.75mm wide and 8mm long implant. Model 3 was bicortication of crestal cortical bone and sinus floor cortical bone using a 5mm wide and 8mm long implant. Model 4 was bicortication of crestal cortical bone and sinus floor cortical bone penetration using a 3.75mm wide and 11mm long implant. Model 5 was maxillary sinus floor elevation by the osteotome technique using a 3.75mm wide and 11mm long implant. Model 6 was maxillary sinus floor elevation by the lateral antrostomy technique using a 3.75mm wide and 11mm long implant. 150N load applied vertically to the central fossa or buccal cusp tip of single-tooth prosthesis was used to evaluate the different models. The conclusions of this computer study are limited by the assumptions involved in the construction of the computer models. Within the scope of this study, the following observations were made. 1. The greatest stress was concentrated on the cortical bone area regardless of the loading conditions(both cental fossa and buccal cusp) 2. Buccal cusp loading caused more stress than central fossa loading in all models. 3. The stress distribution effect was greatest with a wide fixture implant. 4. With bicortication, the stress distribution effect of penetrating the maxillary floor was greater than just engaging the maxillary floor. 5. The stress distribution effect of the lateral antrostomy technique was shown to be greater than the osteotome technique, and the graft cancellous bone had little effect on the stress distribution.
-
자기조직화지도를 이용한 분류문제에서의 결측자료 대체방법 연구
강동주,송주원 한국자료분석학회 2017 Journal of the Korean Data Analysis Society Vol.19 No.3
Self-organizing maps (SOM) (Kohonen, 1990), one of the unsupervised learning neural network models, are used to conduct cluster analysis or visualize high dimensional data in a low dimensional space. Learning vector quantization (LVQ), one of the supervised learning models utilizing the vector quantization method, is an algorithm to adjust decision borders based on a chosen number of nodes for producing bayesian decision borders and usefully applied to discriminant analysis for data with nonlinear decision borders. Here, we conduct a simulation to show that initial nodes of LVQ based on the SOL algorithm performs better in finding optimal nonlinear decision borders than the ones based on the K-mean clustering method. When data with nonlinear decision borders include missing values, imputation can be implemented to fill in missing values. A simulation is conducted to compare the performance of the LVQ method utilizing SOM when missing values are imputed by mean imputation, hotdeck imputation, and a model based imputation. These imputation methods are also applied to impute missing values of glass identification data and misclassification rates of them are compared. Kohonen(1990)에 소개된 자기조직화지도(self-organizing maps, SOM)는 비지도학습(unsupervised learning) 신경망(neural network) 모형의 한 종류로서 고차원 다변량 자료에 대한 군집분석과 저차원 시각화에 사용된다. 학습벡터 양자화(learning vector quantization, LVQ)는 벡터 양자화(vector quantization) 방법을 활용한 지도학습(supervised learning) 모형의 한 종류로서 사전에 정해진 개수의 노드로 구성된 판별경계를 베이즈 판별경계로 미세하게 조정하는 알고리즘이며 비선형 판별경계를 가지는 자료에 대한 판별분석에 유용하게 사용된다. 본 논문에서는 비선형의 판별경계를 구축해야 하는 경우 SOM 방법의 결과노드를 활용하는 것이 K-평균 군집분석의 결과노드를 활용하는 것보다 최적의 판별경계 구성에 유리할 수 있음을 모의실험을 통해 보였다. 또한 결측값이 존재하면서 비선형 판별경계를 가지고 있는 자료에 대해 일반적인 결측값 대체방법인 평균대체, 핫덱대체, 그리고 모형에 근거한 대체 방법으로 대체를 실시하였을 때 SOM을 활용한 LVQ방법의 성능을 모의실험을 통해 알아보고 유리 판별 자료에 각 결측대체방법을 적용하여 판별 결과를 비교하였다.
-
-
-
-
-
연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
