• 전기학회논문지
• /
• 제57권6호
• /
• pp.1058-1062
• /
• 2008

In clinical decision support system(CDSS), unlike rule-based expert method, appropriate data-driven machine learning method can easily provide the information of individual feature(clinical test) for disease classification. However, currently developed methods focus on the improvement of the classification accuracy for diagnosis. With the analysis of feature importance in classification, one may infer the novel clinical test sets which highly differentiate the specific diseases or disease states. In this background, we introduce a novel CDSS that integrate a classifier and feature selection module together. Random forest algorithm is applied for the classifier and the feature importance measure. The system selects the significant clinical tests discriminating the diseases by examining the classification error during backward elimination of the features. The superior performance of random forest algorithm in clinical classification was assessed against artificial neural network and decision tree algorithm by using breast cancer, diabetes and heart disease data in UCI Machine Learning Repository. The test with the same data sets shows that the proposed system can successfully select the significant clinical test set for each disease.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 1999년도 추계학술대회-지능형 정보기술과 미래조직 Information Technology and Future Organization
• /
• pp.365-373
• /
• 1999

Recently, numerous studies have demonstrated that artificial intelligence such as neural networks can be an alternative methodology for classification problems to which traditional statistical methods have long been applied. In building neural network model, the selection of independent and dependent variables should be approached with great care and should be treated as a model construction process. Irrespective of the efficiency of a learning procedure in terms of convergence, generalization and stability, the ultimate performance of the estimator will depend on the relevance of the selected input variables and the quality of the data used. Approaches developed in statistical methods such as correlation analysis and stepwise selection method are often very useful. These methods, however, may not be the optimal ones for the development of neural network models. In this paper, we propose a genetic algorithms approach to find an optimal or near optimal input variables for neural network modeling. The proposed approach is demonstrated by applications to bankruptcy prediction modeling. Our experimental results show that this approach increases overall classification accuracy rate significantly.

• 대한인간공학회지
• /
• 제36권4호
• /
• pp.325-341
• /
• 2017

Objective: The purpose of this study is to suggest the guidelines of operation phases to minimize injuries and musculoskeletal disorders in manual material handling (MMH) tasks through literature reviews. The guidelines are presented as the preparing phase, lifting phase, carrying phase, and lowering phase. Also, we summarized the non-numerical general guidelines for MMH tasks. Background: Manual material handling is still a main cause to musculoskeletal disorders. Method: Procedures of a literature review are classified into database selection, keyword search, title review, abstract review related to literature selection, guideline review and arrangement. A total 48 papers and books were analyzed in detail by title and abstract reviews. Results: In the preparing phase, we suggested the basic conditions in MMH, preparing procedure, clothing and protective equipment, and education. In the lifting and carrying phases, we recommended maximal acceptable weight by frequency and body posture. In the lowering phase, we suggested the lowest weight and safety body postures. Finally, we recommended general guidelines and guideline items for MMH. General guidelines are presented to suggest worker selection, technical education, and work design parts. Conclusion: We suggested the guidelines on the four operation phases of MMH tasks such as preparing, lifting, carrying, and lowering phases. Application: The findings of this study can be utilized as guidelines for proactive recommendations according to workers in MMH tasks.

• 한국항공운항학회지
• /
• 제18권1호
• /
• pp.83-88
• /
• 2010

The study is intended to help to select the pilot candidates with exceptional flight aptitude ability and to investigate the relationship between the results of the scholastic aptitude test and the flight aptitude. It is intended that the research will help to draw recommendations on the relevant fields of the scholastic aptitude test that is to be used to select the high caliber candidates with exceptional flight aptitude ability, to gauge the resulting effectiveness of its application and helping to revise the university's syllabus accordingly. From the study, korean, mathematics, english grade and the flight aptitude test results have all shown to hold mutual relationship and through simple correlation analysis, it was discovered that mathematics and English are the two factors that affect the results of the flight aptitude test, with the extent of its impact graded in descending order of English, mathematics and Korean. Lastly, the logistic regression analysis have discovered that the mathematics grade has significant effect on the classification of the flight aptitude and non aptitude category groups, and English also has significant influence close to the 0.05 p-values. It is believed that should the findings of this study be considered as part of the selection process of the university applicants of the department of aeronautical science, making discovery of candidates of higher quality is expected.

• 한국HCI학회:학술대회논문집
• /
• 한국HCI학회 2009년도 학술대회
• /
• pp.502-507
• /
• 2009

본 논문에서는 실시간 포토 모자이크 생성을 위한 이미지 집합 최적화 기법을 소개한다. 포토 모자이크 기법은 작은 포토 이미지를 사용하여 한 장의 큰 사진을 완성하는 기법으로 이미지를 일정한 영역의 셀로 분할한 후 각 셀 영역을 색상, 모양, 무늬 등에서 적합한 이미지로 대체한다. 큰 사진을 구성하는 이미지에 필요한 다양한 패턴을 얻기 위해서는 많은 양의 포토 타일 이미지가 필요하게 된다. 많은 양의 포토 이미지는 이미지 패턴 검색 시간을 오래 걸리게 하고, 이미지 저장을 위해 많은 리소스를 필요로 한다. 이러한 제약은 실시간 처리나 리소스의 제한이 있는 휴대용 기기에서의 포토 모자이크의 적용을 힘들게 한다. 본 논문에서는 패턴 검색 시간 향상과 메모리 요구 최소화를 위해 유전 알고리즘 활용하여 전체 이미지 데이터베이스에서 가장 특징이 있는 이미지를 선별하여 작은 이미지 집합을 구축한다.

• 한국정밀공학회지
• /
• 제27권11호
• /
• pp.63-71
• /
• 2010

FormFactors is one of the most critical design factors in early development of mobile phone, and proper selection of FormFactors is necessary for smoothness of product development and customer satisfaction. Especially, emergence of mobile convergence added so various functions besides its original call function that importance of selecting FormFactors has increased because multiform FormFactors are needed. However there is problem such as frequent change of decision making in existing process because established process picks out FormFactors sensibly and arbitrarily through idea pull and so forth. We proposed FormFactors selection process by Axiomatic Design, set approach method and block approach method for reasonable and systematical FormFactors selection. First of all, we set the purpose of mobile phone development, and it is examined by Axiomatic Design. FormFactors design matrix is deduced through this process, the numbers of axes and rails are proposed using set approach method, and then patterns of FormFactors are embodied by block approach method. Particularly process application was tried through case study of mobile phone development, and we ensured that new FormFactors can be presented to a designer by systematical verification if change of customer requirements occurs through our process.

• 대한교통학회지
• /
• 제19권5호
• /
• pp.99-108
• /
• 2001

본 논문은 유류저장탱크의 최적입지를 선정하기 위한 GIS와 OR 통합적용기법을 제시하며, P-Median heuristic method와 MIP method가 모두 유류저장탱크의 최적입지 선정 문제에 적용될 수 있음을 실증하였다. 또한, 두 접근 방식에 의해 확인된 결과들은 동일하였다. P-Median heuristic method와 MIP method 를 풀이하는데 필요한 최단거리 매트릭스를 계산해내 기 위해 GIS기반 도로망과 해운네트워크를 구축·결합하였으며, 두 접근방법을 통해 우리나라의 최적 유류저장탱크의 입지를 선정한 결과 권역별로 군산항, 여수항, 부산항, 옥계항이 선정되었다. 위와 같이, 본 연구를 통해 CIS와 P-Median heuristic method, MIP method의 적용은 실증되었으며, 유사한 연구 또한 GIS와 OR의 통합적용을 통해 수행될 수 있을 것이다.

• 시스템엔지니어링학술지
• /
• 제9권1호
• /
• pp.55-63
• /
• 2013

Nuclear power plant site selection is a complex process and its successful completion is a critical milestone in the NPP development cycle. Proper siting of NPP will ensure public health and safety, environmental conservation, reduced project failure risks and a smooth NPP development process among other benefits. The objective of this paper is to demonstrate the application of systems engineering to the problem of NPP siting in Kenya. The siting process demonstrated in this paper includes stakeholder need analysis where stakeholders are identified and their needs concerning NPP site are elicited and converted into system functional requirements. A value model is then developed and potential sites iteratively subjected to three types of criteria i.e. exclusionary criteria, avoidance criteria and suitability criteria. This process is used to identify the candidate sites. An additive value model; multiple objectives Decision Analysis (MODA) is then used to calculate candidate solutions values. The site with the highest solution value score is selected. Sensitivity studies using different criterion weight sets (thereby reflecting different viewpoints) can be conducted to assess their effect on the selection of a preferred site and thereby lend additional credibility to the decision process.

• 한국통신학회논문지
• /
• 제19권2호
• /
• pp.265-275
• /
• 1994

어휘중의성 해소는 음성 인식/생성, 정보 검색, 발뭉치 태킹 등 자연언어 처리에서 가장 기초가 되는 분야 중의 하나이다. 본 논문은 말뭉치로부터 추출된 통계정보를 이용하는 한국어 어휘중의성해소 기법에 대해 기술한다. 이 기법에서는 좀더 정밀한 중의성해소를 위해 품사태그 대신 형태소분석 결과에 해당하는 토큰태그를 사용하고 있다. 본 논문에서 제안한 어휘선택함수는 어미나 조사의 호응 관계등 한국어의 어휘적 특성을 잘 반영하기 때문에 상당히 높은 정확성을 보여준다. 그리고 활용분야에 적합하게 사용될 수 있도록 유일선택 방식과 다중선택 방식이라는 두가지 중의성해소 방식을 지원하고 있다.

• Archives of Pharmacal Research
• /
• 제12권4호
• /
• pp.249-253
• /
• 1989

High producers or blocked mutants of aminoglycoside antibiotic-producing Streptomyces spp. were selected by application of an agar plug method and by culturing individual colonies in broth. The productivities of aminoglycoside antibiotic producing organisms were increased by selection of a high producer from colonies obtained by spreading spores of wild strain, or survived from treatment of a mutagen or from the colonies regenerated from protoplast-formation and cell-wall regenerations. Some mutagen treated colonies lost the ability to produce antibiotics (5-8%). Some A-factor negative and deostreptamine or streptidine negative mutants were obtained by N-methyl-N'-nitro-N-nitrosomethylguanidine (MNNG) treatment. Many of the survivors from the MNNG treatment lost the ability to produce antibiotics. Major colonies produced less amount of antibiotics ; only few survived colonies produced more antibiotics than the parent. Resistance of Streptomyces spp. against the antibiotics produced by itself was also markedly affected by mutagen treatment.