• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권7호
• /
• pp.2683-2702
• /
• 2015

A Content-based Image Retrieval (CBIR) system employs visual features rather than manual annotation of images. The selection of optimal features used in classification of images plays a key role in its performance. Category proliferation problem has a huge impact on performance of systems using Fuzzy Artmap (FAM) classifier. The proposed CBIR system uses a modified version of FAM called Non-Proliferation Fuzzy Artmap (NPFAM). This is developed by introducing significant changes in the learning process and the modified algorithm is evaluated by extensive experiments. Results have proved that NPFAM classifier generates a more compact rule set and performs better than FAM classifier. Accordingly, the CBIR system with NPFAM classifier yields good retrieval.

• 한국IT서비스학회지
• /
• 제4권1호
• /
• pp.117-128
• /
• 2005

When we conduct development of a large-size SI project, it is highly necessary to utilize an appropriate application framework which helps to build a qualified software with high productivity in a short period of time. In this paper, we propose the architecture of a dynamically reconfigurable CBD application integration framework that has been used for developing large-scale e-business applications to achieve high development productivity and maintainability. This Integration framework can easily extend its functionalities, and dynamically change its configuration during run time according to the business category, such as applying interaction patterns among main components in software architecture, rules, policies, and environmental parameters. Dynamic reconfiguration has the feature to make applications be easily customized for changeable requirements. Through our application integration framework, huge sizes of contents can be managed according to the business category as well, by keeping configuration informations and huge volumes of source codes. In order to evaluate out application integration framework in terms of performance criteria, we present experimental results of throughputs from the framework by yielding dynamic configuration without any performance degradation.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2012년도 한국컴퓨터종합학술대회논문집 Vol.39 No.1(B)
• /
• pp.471-473
• /
• 2012

The Category Fluency Test (CFT) is a widely used verbal fluency test. The standard measure of scoring the test is the number of distinct words that a subject generates during the test. Recently, other measures have also been proposed to evaluate performance, such as clustering and switching. In this study, we examine clusters and switches can be assessed using word similarity measures. Based on these measures, we can distinguish between subject groups.

• 반도체디스플레이기술학회지
• /
• 제22권2호
• /
• pp.40-44
• /
• 2023

The defect detection based on deep learning requires minimal loss and high accuracy to pinpoint product defects. In this paper, we confirm the loss rate of deep learning training based on disc-shaped artificial scaffold images. It is intended to compare the performance of Cross-Entropy functions used in object detection algorithms. The model was constructed using normal, defective artificial scaffold images and category cross entropy and sparse category cross entropy. The data was repeatedly learned five times using each loss function. The average loss rate, average accuracy, final loss rate, and final accuracy according to the loss function were confirmed.

• 국제학술발표논문집
• /
• The 4th International Conference on Construction Engineering and Project Management Organized by the University of New South Wales
• /
• pp.292-298
• /
• 2011

In this study we examine Taiwan's overall performance in accordance with sustainable construction by developing an appraisal indicator framework. The framework consists of five layers, from bottom to top: the indicator; the indicator category; the core cluster; the theme; and the overall performance. The procedure for the development of a sustainable construction indicator system is outlined. Finally, a framework consists of 3 themes, 10 core clusters, and 33 indicator categories are established. Following the established framework, 67 proper indicators are selected for each category in the framework, and data of the 53 indicators are collected respectively from a nation's statistical databank in Taiwan. Sustainable construction index aggregated step-by-step from the indicators, the indicator categories, the core clusters and the themes is computed to assess Taiwan's progress in sustainable construction. The preliminary results and the discussion are reported.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제8권9호
• /
• pp.363-372
• /
• 2019

시퀀스란 두 항목 간의 순서가 존재하는 데이터를 말하며, 고객 한 명이 구매한 상품들이 나열된 구매이력 데이터는 대표적인 시퀀스 데이터 중 하나이다. 일반적으로 모든 상품은 대분류/ 중분류/ 소분류와 같은 상품 분류 체계를 가지며, 서로 다른 상품이더라도 비슷하다면 그 특성에 따라 동일한 범주로 분류된다. 따라서 본 논문에서는 두 구매이력 시퀀스 비교 시 상품의 구매 순서를 고려할 뿐만 아니라, 비교하고자 하는 두 상품이 다르더라도 서로 동일한 상품 군에 속한다면 더 높은 유사도를 부여하여 계산한다. 특히 구매이력 시퀀스 유사도 계산 성능에 직접적인 영향을 미치는 시퀀스 유사도 측정 방법을 선택하기 위해 본 연구에서는 대표적인 시퀀스 간 유사도 측정 방법인 레벤슈타인 거리, 동적 타임 워핑 거리, 니들만-브니쉬 유사도의 성능을 비교하였으며, 항목간의 계층구조도 반영하여 계산하도록 확장하였다. 기존의 유사도 측정 방법의 경우 시퀀스 내 상품 비교 시 상품의 일치 유무에 따라 단순히 0 또는 1의 값을 부여하여 계산한다. 하지만 제안 방법의 경우 서로 다른 상품이더라도 두 상품 간의 연관정도를 다르게 부여하기 위하여 상품 분류 트리를 사용하여 0에서 1 사이의 값을 가지도록 세분화하였다. 실험을 통해 세 알고리즘에 제안 방법을 적용한 경우 기존 방법에 비하여 구매이력 시퀀스 간의 유사도를 더 정확히 측정함을 확인하였다. 또한 정확성 측정 비교 실험을 통해 동적 타임 워핑 유사도가 다른 두 유사도 측정 방법에 비하여 시퀀스 내 상품의 연관 정도를 고려할 뿐만 아니라 두 시퀀스의 길이가 다른 경우에도 좋은 성능을 보였기 때문에 구매이력 데이터에서 시퀀스 간의 유사도 비교 시 가장 적합한 측정 방법임을 확인하였다.

• 한국주거학회논문집
• /
• 제17권2호
• /
• pp.77-85
• /
• 2006

This study aims to propose the comprehensive evaluation of performance multi-family housing. Most of the existing studies for the reproduction have evaluated by the safety of structure. Therefore those pass over the functional factors which is important for the evaluation of residential qualities. This study represents that the reproduction renovation should not be secured the safety of structure but established the quality of lift. And questionnaire asking expert's opinion also supplemented. As a result, the item for examination and the weight for evaluation are suggested in functional deterioration. The results of this study are summarized as follows: 1. Existent redevelopment criteria are not appropriate to decide the environmental comfort, the vital function of housing and the residential management finance. 2. The future criteria have to evaluate the existing environmental comfort, be reconstructed to put first general deterioration. 3. Evaluation elements was extracted throughout the document research, evaluation elements analysis system was separated in residential area, suburb of dwellings, community, and housing complexity environmental assesment, 4. Each category of influence ratio was extracted to assess total deterioration throughout the expert question survey. 5. Each category analysed (1) $0{\sim}30%$: function is good (unnecessary to improve) (2) $31{\sim}40%$: average (necessary of partial improvement) (3) $41{\sim}59%$: bad (necessary of total improvement) (4) $60{\sim}100%$: the worst (unable to improve throughout the repair)

• Journal of Communications and Networks
• /
• 제13권2호
• /
• pp.149-159
• /
• 2011

In this paper, we study the problem of how to design a medical-grade wireless local area network (WLAN) for healthcare facilities. First, unlike the IEEE 802.11e MAC, which categorizes traffic primarily by their delay constraints, we prioritize medical applications according to their medical urgency. Second, we propose a mechanism that can guarantee absolute priority to each traffic category, which is critical for medical-grade quality of service (QoS), while the conventional 802.11e MAC only provides relative priority to each traffic category. Based on absolute priority, we focus on the performance of real-time patient monitoring applications, and derive the optimal contention window size that can significantly improve the throughput performance. Finally, for proper performance evaluation from a medical viewpoint, we introduce the weighted diagnostic distortion (WDD) as a medical QoS metric to effectively measure the medical diagnosability by extracting the main diagnostic features of medical signal. Our simulation result shows that the proposed mechanism, together with medical categorization using absolute priority, can significantly improve the medical-grade QoS performance over the conventional IEEE 802.11e MAC.

• 한국문헌정보학회지
• /
• 제57권2호
• /
• pp.435-452
• /
• 2023

이 연구는 딥러닝 기법의 전이학습 모형인 BERT를 이용하여 주제명의 자동 분류를 실험하고 그 성능을 평가하였으며, 더 나아가 주제명이 부여된 KDC 분류체계와 주제명의 범주 유형에 따른 성능을 분석하였다. 실험 데이터는 국가서지를 이용하여 주제명의 부여 횟수에 따라 6개의 데이터셋을 구축하고 분류 자질로 서명을 이용하였다. 그 결과, 분류 성능으로 3,506개의 주제명이 포함된 데이터셋(레코드 1,539,076건)에서 마이크로 F1과 매크로 F1 척도가 각각 0.6059와 0.5626 값을 보였다. 또한 KDC 분류체계에 따른 분류 성능은 총류, 자연과학, 기술과학, 그리고 언어 분야에서 좋은 성능을 보이며 종교와 예술 분야는 낮은 성능을 보였다. 주제명의 범주 유형에 따른 성능은 '식물', '법률명', '상품명'이 높은 성능을 보인 반면, '국보/보물' 유형의 주제명에서 낮은 성능을 보였다. 다수의 주제명을 포함하는 데이터셋으로 갈수록 분류기가 주제명을 제대로 부여하지 못하는 비율이 늘어나 최종 성능의 하락을 가져오기 때문에, 저빈도 주제명에 대한 분류 성능을 높이기 위한 개선방안이 필요하다.

• 한국군사과학기술학회지
• /
• 제11권2호
• /
• pp.88-94
• /
• 2008

It is very important for weapon systems to manage the logistic support for the better combat readiness. Therefore, in this paper we review the issues of logistic support and suggest alternatives to effectively manage the logistic support for weapon systems by using performance based logistics, especially during operations and support phase.