• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2017.04a
• /
• pp.166-166
• /
• 2017

모돈의 건강 상태를 정량 지수화 하기 위한 연구를 수행 중이다. 지제이상, 섭식 불량, 수면 패턴 등의 운동 특성 분석을 위하여 복수의 초음파 센서를 이용하였다. 시계열 계측 신호를 분석하여 정량 지수화를 수행하는 과정에서 주파수 도메인 분석을 시도하였다. 이 과정에서 주파수 도메인의 분해능에 따른 편차 극복을 위한 비선형 모델링을 수행하였다. 또한 인접한 시계열 데이터 구간 간의 상관성 분석이 가능하면 대용량 데이터의 실시간 처리로 인한 지연 시간 극복 및 기대되는 예후에 대한 조기 진단이 가능할 것이다. 본 연구에서는 구글에서 제공하는 Tensorflow와 NVIDIA에서 제공하는 CUDA 엔진을 동시 적용한 심층 학습 시스템을 이용하였다. 전 처리를 위하여 주파수 분해능 (2분, 3분, 5분, 7분, 11분, 13분, 17분, 19분)에 따른 데이터 집합을 1단계로 두고, 상위 10 순위 안에 드는 파워 스펙트럼 밀도의 크기를 2단계로 하여, 총 2~10개의 입력 노드를 순차적으로 선정하였고, 동일한 방식으로 인접한 시계열의 파워 스펙터럼 밀도를 순위를 변화시켜 지정하였다. 대표적인 심층학습 모델인 Softmax regression with a multilayer convolutional network를 이용하여 Recursive feature selection 경우의 수를 $8{\times}9{\times}9$로 총 648 가지 선정하고, Epoch는 10,000회로 지정하였다. Calibration 모델링의 경우 Cost function이 10% 이하인 경우 해당 경우의 학습을 중단하였으며, 모델 간 상호 교차 검증을 수행하기 위하여 $_8C_2{\times}_8C_2{\times}_8C_2$ 경우의 수에 대한 Verification test를 수행하였다. Calibration 과정 상 모든 경우에 대하여 10% 이하의 Cost function 값을 보였으나, 검증 테스트 과정에서 모든 경우에 대하여 $r^2$ < 0.5 인 결정 계수 값이 나타났다. 단적으로 심층학습 모델의 과도한 적합(Over fitting) 방식의 한계를 보인 것이라 판단할 수 있다. 적합한 Feature selection 및 심층 학습 모델에 대한 지속적이고 추가적인 고려를 통해 과도적합을 해소함과 동시에 실효적이고 활용 가능한 Classification을 위한 입, 출력 노드 단의 전후 Indexing, Quantization에 대한 고려가 필요할 것이다. 이를 통해 모돈 생체 정보 정량화를 위한 지능형 현장 진단 기술 연구를 지속할 것이다.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.46 no.10
• /
• pp.806-813
• /
• 2018

The fracture of a turbine blade of aerospace engine is presented. Although there are a lot of causes and failure modes in blades, the main failure modes are two ways that fracture and fatigue. Degradation of blade material affects most failure modes. Total propagation of failure in this study specifies failure of fracture type. Some section appears fatigue mode. Especially since this study describes analysis of failure for blade in high temperature, it can be a case in point. Analysed blade is Ni super alloy. Investigations of blade are visual inspection, material, microstructure, high temperature stress rupture creep test, analysis and fracture surface, etc. The root cause for fracture was stress rupture due to abnormal thermal environment. Thermal property of Ni super alloy is excellent but if each chemical composition of alloys are different due to change mechanical properties, selection of material is very important.

• Journal of Navigation and Port Research
• /
• v.29 no.6 s.102
• /
• pp.529-536
• /
• 2005

Recently the building of ultra large container ship are discussed among ship building companies and ship operating company who have a tendency to pursue the advantage of large scale of economy. These tendency will be continued for the time being, if ship-building skill and economical efficiency are available. As the enlargement of container ship size becomes hot issues in ship-building markets, the needs for re-arrangement of accommodation house in large container ship are proposed carefully in some researches. This study examined economical efficiency of re-arrangement of accommodation house in ultra large container ship. The separation between accommodation and engine room is proposed through out drawing works in initial design stage and we examined the merits and demerits of the separation in the view of economical efficiency. The RFR(Required Freight Rate) is considered as the objective function to evaluate the re-designed vessel. The economical benefits are analyzed in the view of ship operator and shipyard respectively.

• Korean Chemical Engineering Research
• /
• v.48 no.3
• /
• pp.375-381
• /
• 2010

Cetane number of diesel fuel for compression ignition engine is one of main properties for fuel ignition quality. Recently the cetane index has been replaced the cetane number in order to resolve the disadvantage of CFR engine test, but these two value have slightly difference values due to addition of various additives. In this study, we analyzed the cetane number, derived cetane number and cetane index for diesel fuel which was blended with various ratios of biodiesel, kerosene and cetane improver as additives. As a result, Cetane number showed the similar value with derived cetane number, but cetane index showed quite different value with cetane number when biodiesel and cetane improver were used as additives.

• Annual Conference on Human and Language Technology
• /
• 1999.10d
• /
• pp.60-75
• /
• 1999

품사 태깅은 자연 언어 처리의 가장 기본이 되는 부분으로 상위 자연 언어 처리 부분인 구문 분석, 의미 분석의 전처리로 사용되고, 독립된 응용으로 언어의 정보를 추출하거나 정보 검색 등의 응용에 사용되어 진다. 품사 태깅은 크게 통계에 기반한 방법, 규칙에 기반한 방법, 이 둘을 모두 이용하는 혼합형 방법 등으로 나누어 연구되고 있다. 포항공대 자연언어처리 연구실의 자연 언어 처리 엔진(SKOPE)의 품사 태깅 시스템 POSTAG는 미등록어 추정이 강화된 혼합형 품사 태깅 시스템이다 본 시스템은 형태소 분석기, 통계적 품사 태거, 에러 수정 규칙 후처리기로 구성되어 있다. 이들은 각각 단순히 직렬 연결되어 있는 것이 아니라 형태소 접속 테이블을 기준으로 분석 과정에서 형태소 접속 그래프를 생성하고 처리하면서 상호 밀접한 연관을 가진다. 그리고, 미등록어용 패턴사전에 의해 등록어와 동일한 방법으로 미등록어를 처리함으로써 효율적이고 강건한 품사 태깅을 한다. 한편, POSTAG에서 사용되는 태그세트와 한국전자통신연구원(ETRI)의 표준 태그세트 간에 양방향으로 태그세트 매핑을 함으로써, 표준 태그세트로 태깅된 코퍼스로부터 POSTAC를 위한 대용량 학습자료를 얻고 POSTAG에서 두 가지 태그세트로 품사 태깅 결과 출력이 가능하다. 본 시스템은 MATEC '99'에서 제공된 30000어절에 대하여 표준 태그세트로 출력한 결과 95%의 형태소단위 정확률을 보였으며, 태그세트 매핑을 제외한 POSTAG의 품사 태깅 결과 97%의 정확률을 보였다.

• Journal of the Korea Society of Computer and Information
• /
• v.7 no.4
• /
• pp.141-149
• /
• 2002

Recently, a number of network systems are developed rapidly and network architectures are more complex than before, and a policy-based network management should be used in network system. Especially, a new paradigm that policy-based network management can be applied for the network security is raised. A security policy server in the management layer can generate new policy, delete. update the existing policy and decide the policy when security policy is requested. The security server needs to analyze and manage the alert message received from server Policy enforcement system in the enforcement layer for the available information. In this paper, we implement an alert analyzer that analyze the stored alert data for making of security policy efficiently in framework of the policy-based network security management. We also propose a data mining system for the analysis of alert data The implemented mining system supports alert analyzer and the high level analyzer efficiently for the security.

• The Journal of Korean Association of Computer Education
• /
• v.16 no.1
• /
• pp.97-110
• /
• 2013

The purpose of this study is to analyze the needs of teachers as main users to make suggestions for active use of educational information sharing system. The study sets three research tasks about the problems of the current system, improvements, and desired content from the perspective of teachers. To collect specific opinions, we conducted an open questionnaire survey and made a qualitative analysis using the NVivo 10 program. The present system has such problems as functional issues, lack of data reliability, systemic problems for sharing and use. We suggest that in order to enhance Web site services, authoring tools for editing contents and new search engines with customized search function need to be provided. In addition, the open market service system has to be established, so that teachers can actively create, share, and distribute to meet the various needs of teachers. Finally, to promote free access to the system, copyright issues need to be resolved to prevent the potential digital piracy.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.27 no.2
• /
• pp.81-94
• /
• 2016

This study analyzes the problems in reading program of public libraries and provides some suggestions for the program development and management. For this analysis, I sampled and examined 350 news articles to reading programs done in the public libraries between 2013 and 2015. The result shows that 41.4% of the collected articles were the news about the reading programs in Seoul Metropolitan areas. While the articles about simple reading programs occupied 69.9%, the mixed activity programs reached a significant amount in total. In addition, 51.3% of program participants was under the elementary children and their parents, but only 9.0% program participants came from low socioeconomic status classes. 86 reading programs (28.4%) were designed for one-day events without a long-term policy. Based on such analyses, this study provides three suggestions: utilizing public relation teams and communicating mass media to advertize reading programs to local residents, developing systematic reading programs in users' needs, managing the reading programs for all kinds of local residents.

• Journal of the Korean Data and Information Science Society
• /
• v.26 no.5
• /
• pp.1155-1166
• /
• 2015

It is almost impossible to store or analyze big data increasing exponentially with traditional technologies, so Hadoop is a new technology to make that possible. In recent R is using as an engine for big data analysis based on distributed processing with Hadoop technology. With RHadoop that integrates R and Hadoop environment, we implemented parallel multiple regression analysis with various data sizes of actual data and simulated data. Experimental results showed our RHadoop system was faster as the number of data nodes increases. We also compared the performance of our RHadoop with lm function and biglm packages available on bigmemory. The results showed that our RHadoop was faster than other packages owing to paralleling processing with increasing the number of map tasks as the size of data increases.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.6
• /
• pp.1458-1467
• /
• 1999

Existing index structures such as extendable hashing and B+-tree do not support partial string searches perfectly. The inverted file method and the signature file method that are used in the web retrieval engine also have problems that they do not provide partial string searches and suffer from serious retrieval performance degradation respectively. In this paper, we propose an efficient index method that supports partial string searches and achieves good retrieval performance. The proposed index method is based on the Inverted file structure. It constructs the index file with patterns that result from dividing terms by two syllables to support partial string searches. We analyze the characteristics of our proposed method through simulation experiments using wide range of parameter values. We analyze the derive analytic performance evaluation models of the existing inverted file method, signature file method and the proposed index method in terms of retrieval time and storage overhead. We show through performance comparison based on analytic models that the proposed method significantly improves retrieval performance over the existing method.