• 정보과학회 논문지
• /
• 제43권8호
• /
• pp.902-909
• /
• 2016

질의응답 시스템은 사용자의 질문에 대한 답을 찾아주는 시스템으로, 기존의 검색엔진이 사용자의 질의에 대해 관련된 문서의 링크만을 찾아주는 반면 질문에 대한 최종적인 답을 찾아준다는 차이점이 있다. 특정 분야에 국한되지 않고 다양한 질문을 처리해주는 오픈 도메인 질의응답 시스템에 필요한 연구들이 최근 자연어 처리, 인공지능, 데이터 마이닝 등 학계의 다양한 분야들에서 뜨거운 관심을 받고 있다. 하지만 관련 연구에서는 학습 데이터에는 없었던 단어들이 질문에 대한 정확한 답과 유사한 오답을 구별해내는데 결정적인 역할을 할 수 있음에도, 이러한 처음 보는 단어들을 모두 단일 토큰으로 치환해버리는 문제가 있다. 본 논문에서는 문맥 정보를 통해 이러한 모르는 단어에 대한 벡터를 계산하는 방법을 제안한다. 그리고 역문헌빈도 가중치를 활용하여 문맥정보를 더 효율적으로 처리하는 모델을 제안한다. 또한 풍부한 실험을 통해 질의응답 시스템의 모델 학습 속도 및 정확성이 기존 연구에 비해 향상됨을 확인하였다.

• 디지털융복합연구
• /
• 제14권1호
• /
• pp.53-63
• /
• 2016

본 연구는 사회서비스에서 비중이 급격히 확대되고 있는 지역사회서비스투자사업에 초점을 맞추고 의미 있는 발전을 위해 필요한 논의가 무엇인가를 찾고자 하였다. 산물분석 틀을 이용하여 지역사회서비스투자사업의 내용을 분석하고 사업을 구성하는 사회적 선택의 근거가 되는 가치를 축으로 분석하였다. 지역사회서비스투자사업의 내용과 가치 분석을 토대로 도출한 정책적 쟁점은 다음과 같다. 첫째, 지역사회서비스투자사업은 지금보다 더 '지역'에 초점을 맞출 필요가 있다. 둘째, 시민의 개별성에 초점을 맞추는 사회서비스가 더 중요해질수록 주민참여에 기반한 지역사회서비스투자사업의 '지속가능성'에 대해 생각할 필요가 있다. 셋째, 지역사회의 '공동체 형성'과 관련하여 지역사회서비스투자사업의 목표 수정에 대해 논의할 필요가 있다. 이를 위해 지역주민의 의사결정 통로 부재, 소득 기준 대상 선정, 사업명칭에서 '투자'라는 용어의 적절성에 대해 재검토가 필요하다.

• 한국인터넷방송통신학회논문지
• /
• 제10권3호
• /
• pp.91-95
• /
• 2010

검색엔진을 사용하는 이용자의 정보 즉 선호도에 따른 지속적인 피드백으로 검색 결과의 랭킹을 향상시켜 유연한 검색이 가능하게 하는 방법에는 학습된 인공 신경망을 이용한다. 인공 신경망 학습은 신경망이 여러 다른 검색어로 학습된 후 다른 사용자들이 과거에 실제 검색했던 결과를 좀 더 반영하기 위한 것이다. 가중치의 지속적인 변경을 위해서는 네트워크에서 역방향으로 움직이면서 가중치를 변경하는 역전파 알고리즘을 이용하여 학습한다. 그러나 이러한 학습은 초기에는 훈련데이터에 적합한 성능을 보이나 학습의 횟수가 증가할수록 점점 과대적합되는 것을 알 수 있다. 따라서 본 논문에서는 최적화해야 할 개체가 많을 때 강한 장점을 가지고 있는 유전자 알고리즘을 적용하여 검색어에 관련성이 높은 페이지들 유연하게 랭킹하기 위해 URL리스트를 개체로 랜덤으로 선택하여 학습하는 기법을 제안한다.

• 전기학회논문지
• /
• 제60권1호
• /
• pp.167-173
• /
• 2011

Most of the terminologies used in medical domain is not intended to be applied directly in clinical setting but is developed to integrate the terms by defining the reference terminology or concept relations between the terms. Therefore, it is needed to develop the subsets of the terminology which classify categories properly for the purpose of use and extract and organize terms with high utility based on the classified categories in order to utilize the clinical terms conveniently as well as efficiently. Moreover, it is also necessary to develop and upgrade the terminology constantly to meet user's new demand by changing or correcting the system. This study has developed a mapping tool that allows accurate expression and interpretation of clinical terms used for medical records in electronic medical records system and can furthermore secure semantic interoperability among the terms used in the medical information model and generate common terms as well. The system is designed to execute both 1:1 and N:M mapping between the concepts of terms at a time and search for and compare various terms at a time, too. Also, in order to enhance work consistency and work reliability between the task performers, it allows work in parallel and the observation of work processes. Since it is developed with Java, it adds new terms in the form of plug-in to be used. It also reinforce database access security with Remote Method Invocation (RMI). This research still has tasks to be done such as complementing and refining and also establishing management procedures for registered data. However, it will be effectively used to reduce the time and expenses to generate terms in each of the medical institutions and improve the quality of medicine by providing consistent concepts and representative terms for the terminologies used for medical records and inducing proper selection of the terms according to their meaning.

• 대한인간공학회지
• /
• 제29권6호
• /
• pp.831-841
• /
• 2010

Usability evaluation of physical products involves characterizing complex physical interactions between humans and products. Human models known as manikins have been widely utilized as usability evaluation tools for automobile interior package design. When combined with computer-aided design software programs, such manikins can be used to simulate driving postures and evaluate driver-interior fits early in the design process, and therefore, may greatly facilitate achieving high-quality design in a cost-efficient manner. The purpose of this study was to define a set of manikins for designing automobile interior packages for the South Korean male population. These manikins were conceptualized as "boundary" manikins, which represent individuals lacking in certain physical capacities or having usability-related issues (e.g., an individual with the 5th percentile forward reach capability, an individual with the 95th percentile shoulder width). Such boundary manikins can serve as an efficient tool for determining if an automobile interior design accommodates the majority of the population. The boundary manikins were selected from the large sample of Korean males whose anthropometric dimensions were described in the recent Size Korea anthropometric database. For each male in the database, his comfortable driving posture was represented using a kinematic body linkage model and various physical capacity measured and usability-related characteristics relevant to driver accommodation were evaluated. For each such measure, a boundary manikin was selected among the Korean males. The manikins defined in this study are expected to serve as tools for ergonomic design of automobile interior packages. The manikin selection method developed in this study was implemented as a generic software program useful for various product design applications.

• 산업경영시스템학회지
• /
• 제38권3호
• /
• pp.56-63
• /
• 2015

Several fields of science have demanded large-scale workflow support, which requires thousands of CPU cores or more. In order to support such large-scale scientific workflows, high capacity parallel systems such as supercomputers are widely used. In order to increase the utilization of these systems, most schedulers use backfilling policy: Small jobs are moved ahead to fill in holes in the schedule when large jobs do not delay. Since an estimate of the runtime is necessary for backfilling, most parallel systems use user's estimated runtime. However, it is found to be extremely inaccurate because users overestimate their jobs. Therefore, in this paper, we propose a novel system for the runtime prediction based on workload-aware clustering with the goal of improving prediction performance. The proposed method for runtime prediction of parallel applications consists of three main phases. First, a feature selection based on factor analysis is performed to identify important input features. Then, it performs a clustering analysis of history data based on self-organizing map which is followed by hierarchical clustering for finding the clustering boundaries from the weight vectors. Finally, prediction models are constructed using support vector regression with the clustered workload data. Multiple prediction models for each clustered data pattern can reduce the error rate compared with a single model for the whole data pattern. In the experiments, we use workload logs on parallel systems (i.e., iPSC, LANL-CM5, SDSC-Par95, SDSC-Par96, and CTC-SP2) to evaluate the effectiveness of our approach. Comparing with other techniques, experimental results show that the proposed method improves the accuracy up to 69.08%.

• 대한지역사회영양학회지
• /
• 제3권4호
• /
• pp.655-667
• /
• 1998

The Computer-assisted Hazard Analysis and Critical Control Point(HACCP) program has been developed for a systematic implementation of HACCP principles in identifying, assessing and controlling hazards in institutional foodservics operations. The HACCP-based sanitation evaluation tool has been developed, based on the results of the computerized assisted HACCP program in 4 service sites of C contracted foodservice company, including 2 general hospitals with 650-beds, one office operation of 400 meals per day, and one factory foodservice of 1,000 meals per day. All database files and processing programs were created by using Unify Vision tool with Windows 95 of user environments. The results of this study can be summarized as follows : 1. This program consists of the pre-stage for HACCP study and the implementation stage of the HACCP system. 1) The pre-stage for HACCP study includes the selection of menu items, the development of the HACCP recipe, the construction of product flow diagrams, and printing the HACCP recipes and product flow diagrams. 2) The implementation of the HACCP system includes the identification of microbiological hazards, the determination of critical control points based on the decision tree base files. 3) The HACCP-based sanitation evaluation tool consisted of 3 dimensions of time-temperature relationship, personal hygiene, and equipment-facility sanitation. The Cronbach's alphas calculation indicated that the tool was reliable. The results showed that the focus groups rated the mean of importance in time-temperature relationship, personal hygiene, and equipment-facility sanitation as 4.57, 4.59 and 4.55 respectively. Based on the results, this HACCP-based sanitation evaluation tool was considered as an effective tool for assuring product quality. This program will assist foodservice managers to encourage a standardized approach in the HACCP study and to maintain a systematic approach for ensuring that the HACCP principles are applied correctly.

• Journal of Korean Neurosurgical Society
• /
• 제39권6호
• /
• pp.471-474
• /
• 2006

Objective : Neuroimaging data are of paramount importance in making correct diagnosis. We herein evaluate the clinical usefulness of image transfer using cellular phones to facilitate neurological diagnosis and decision-making. Methods : Selected images from CT, MRI scans, and plain films obtained from 50 neurosurgical patients were transferred by cellular phones. A cellular phone with a built-in 1,300,000-pixel digital camera was used to capture and send the images. A cellular phone with a 262,000 color thin-film transistor liquid crystal display was used to receive the images. Communication between both cellular phones was operated by the same wireless protocol and the same wireless internet service. We compared the concordance of diagnoses and treatment plans between a house staff who could review full-scale original films and a consultant who could only review transferred images. These finding were later analyzed by a third observer. Results : The mean time of complete transfer was $2{\sim}3\;minutes$. The quality of all images received was good enough to make precise diagnosis and to select treatment options. Transferred images were helpful in making correct diagnosis and decision making in 49/50 [98%] cases. Discordant result was caused in one patient by improper selection of images by the house staff. Conclusion : The cellular phone system was useful for image transfer and delivery patient's information, leading to earlier diagnosis and initiation of treatment. This usefulness was due to sufficient resolution of the built-in camera and the TFT-LCD, the user-friendly features of the devices, and their low cost.

• 품질경영학회지
• /
• 제49권3호
• /
• pp.421-431
• /
• 2021

Purpose: For improving outgoing quality, this study presents a novel sampling framework based on predictive analytics. Methods: The proposed framework is composed of three steps. The first step is the variable selection. The knowledge-based and data-driven approaches are employed to select important variables. The second step is the model learning. In this step, we consider the supervised classification methods, the anomaly detection methods, and the rule-based methods. The applying model is the third step. This step includes the all processes to be enabled on real-time prediction. Each prediction model classifies a product as a target sample or random sample. Thereafter intensive quality inspections are executed on the specified target samples. Results: The inspection data of three Samsung products (mobile, TV, refrigerator) are used to check functional defects in the product by utilizing the proposed method. The results demonstrate that using target sampling is more effective and efficient than random sampling. Conclusion: The results of this paper show that the proposed method can efficiently detect products that have the possibilities of user's defect in the lot. Additionally our study can guide practitioners on how to easily detect defective products using stratified sampling

• 기록학연구
• /
• 제10호
• /
• pp.226-273
• /
• 2004

The characteristics of the records management systems of the Japanese colonial government can be summarized as follows. First, the Government General adopted a "decentralized retention" of public records. The Government General did not establish its own archives for central preservation of permanent public records. Colonial agencies established its own records office and the records office managed the records its agency created. Secret records and police records were exception. They were retained by the Secretary Office of the General Affairs Division and Police Division of the Chosun Government General respectively. Second, filing systems and retention periods of the public records followed the hierarchic structure of organization. In the headquarter of the government, records were filed by a "bureau-division-activity-file" classification system and a retention period of a file was given automatically by each unit the file belonged. A closed and cut-off file was retained and arranged according to its creating unit, creating year, and retention period. The filing system was easy to use once the filing system was established well, but to make it work effectively changes in activities and organizations should be on a reflected regular basis. It had an advantageous effect that permanent records could be preserved in a unified way throughout the organization. However, it is very critical to determine the permanent records in a professional way. Selection of the permanent records should be done professionally and in a historical perspective. Otherwise, the records retained as permanent records were not the records having an enduring value. And that was not done by the colonial government. Third, classification and scheduling of records were carried out by a creating division, rather than by the Records Office, mostly from the 1920s. Compilation of the records was also done by the creating agency. It implies that the records management lacked the professionalism. In conclusion, the records management system of the Chosun Government General wes nither modern nor user oriented. It managed the records for solely administrative purpose, i.e. effective colonial rule. The legacy of the colonial records system still exists in the public records system in Korean government. One should criticize the lack of will and efforts to modernize the public records system since the establishment of the Korean government while should reflect the historical origins of the records system in Korea.