• Journal of Computing Science and Engineering
• /
• 제10권3호
• /
• pp.85-94
• /
• 2016

We propose a model-based method for information alignment using educational standards as a case study. Discrepancies and inconsistencies in educational standards across different states/cities hinder the retrieval and sharing of educational resources. Unlike existing educational standards alignment systems that only give binary judgments (either "aligned" or "not-aligned"), our proposed system classifies each pair of educational standard statements in one of seven levels of alignments: Strongly Fully-aligned, Weakly Fully-aligned, Partially-$aligned^{***}$, Partially-$aligned^{**}$, Partially-$aligned^*$, Poorly-aligned, and Not-aligned. Such a 7-level categorization extends the notion of binary alignment and provides a finer-grained system for comparing educational standards that can broaden categories of resource discovery and retrieval. This study continues our previous use of mathematics education as a domain, because of its generally unambiguous concepts. We adopt a materialization pattern (MP) model developed in our earlier work to represent each standard statement as a verb-phrase graph and a noun-phrase graph; we align a pair of statements using graph matching based on Bloom's Taxonomy, WordNet, and taxonomy of mathematics concepts. Our experiments on data sets of mathematics educational standards show that our proposed system can provide alignment results with a high degree of agreement with domain expert's judgments.

• Journal of Electrical Engineering and Technology
• /
• 제10권3호
• /
• pp.1264-1274
• /
• 2015

Human action recognition received many interest in the computer vision community. Most of the existing methods focus on either construct robust descriptor from the temporal domain, or computational method to exploit the discriminative power of the descriptor. In this paper we explore the idea of using local action attributes to form an action descriptor, where an action is no longer characterized with the motion changes in the temporal domain but the local semantic description of the action. We propose an novel framework where introduces local action attributes to represent an action for the final human action categorization. The local action attributes are defined for each body part which are independent from the global action. The resulting attribute descriptor is used to jointly model human action to achieve robust performance. In addition, we conduct some study on the impact of using body local and global low-level feature for the aforementioned attributes. Experiments on the KTH dataset and the MV-TJU dataset show that our local action attribute based descriptor improve action recognition performance.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2008년도 추계학술발표대회
• /
• pp.702-705
• /
• 2008

인터넷 사용자 증가에 따라 검색 엔진의 사용 또한 급격히 늘어나고 있는 추세이다. 국내외 다양한 검색 엔진들이 존재하지만 대부분의 자료들이 기본적인 카테고리별로 링크 횟수나 키워드 빈발 횟수에 따라 정렬이 되어 있다. 그러므로 사용자들은 수동적으로 정렬된 도메인들을 따라 가는 실정이다. 본 논문에서는 수동적인 서비스가 아닌 능동적인 서비스에 중점을 둔다. 특정 카테고리 내에서 접속한 사용자에게 최근 시점을 기준으로 가장 빈번하게 접속된 도메인 정보를 제공하여 시간의 단축과 유용한 서비스를 받도록 한다. 본 논문의 서비스 모델은 인터넷 사용자의 로그 데이터베이스와 도메인 데이터베이스를 기반으로 한다. 본 논문에서 제안하는 카테고리 분류 기법으로 두 데이터베이스를 통합하고 정제한다. 정제된 데이터들은 최종적으로 순차 패턴 마이닝 기법에 의해 최종 빈발 패턴을 추출 하게 되고 특정 카테고리에 접속한 사용자에게 도메인 형태로 변환 되어 서비스 하게 된다.

• 한국전자거래학회지
• /
• 제14권4호
• /
• pp.59-73
• /
• 2009

키워드 추출은 정보검색, 문서 분류, 요약, 주제탐지 등의 텍스트 마이닝 분야에서 기반이 되는 기술이다. 대용량 전자문서로부터 추출된 키워드들은 텍스트 마이닝을 위한 중요 속성으로 활용되어 문서 브라우징, 주제탐지, 자동분류, 정보검색 시스템 등의 성능을 높이는데 기여한다. 본 논문에서는 인터넷 포털 사이트에 게재되는 대용량 뉴스문서집합을 대상으로 키워드 추출을 수행하여 분야별 주제를 제시할 수 있는 키워드를 추출하는 새로운 기법을 제안한다. 기본적으로 키워드 추출을 위해 기존 TF-IDF 모델을 고찰, 이것의 6가지 변형식을고안하여 이를 기반으로 각 분야별 후보 키워드를 추출한다. 또한 분야별로 추출된 단어들의 분야간 교차비교분석을 통해 불용어 수준의 의미 없는 단어를 제거함으로써 그 성능을 높인다. 제안 기법의 효용성을 입증하기 위해 한글 뉴스 기사 문서에서 추출한 키워드의 질을 비교하였으며, 또한 주제 변화를 탐지하기 위해 시간에 따른 키워드 집합의 변화를 보인다.

• 한국융합학회논문지
• /
• 제13권3호
• /
• pp.1-9
• /
• 2022

UML(Unified Modeling Language) 클래스 다이어그램은 시스템의 정적인 측면을 표현하며 분석 및 설계부터 문서화, 테스팅까지 사용된다. 클래스 다이어그램을 이용한 모델링이 소프트웨어 개발에 있어 필수적이지만, 경험이 많지 않은 모델러에게 쉽지 않은 작업이다. 도메인 카테고리별로 분류된 클래스 다이어그램 데이터 세트가 제공된다면, 모델링 작업의 생산성을 높일 수 있을 것이다. 본 논문은 클래스 다이어그램 이미지 데이터를 구축하기 위한 자동 분류 기술을 제공한다. 추가 정보 없이 단지 UML 클래스 다이어그램 이미지를 식별하고 도메인 카테고리에 따라 자동 분류한다. 먼저, 웹상에서 수집된 이미지들이 UML 클래스 다이어그램 이미지인지 여부를 판단한다. 그리고, 식별된 클래스 다이어그램 이미지에서 클래스 이름을 추출하여 도메인 카테고리에 따라 분류한다. 제안된 분류 모델은 정밀도, 재현율, F1점수, 정확도에서 각각 100.00%, 95.59%, 97.74%, 97.77%를 달성했으며, 카테고리별 분류에 대한 정확도는 81.1%와 95.2% 사이에 분포한다. 해당 실험에 사용된 클래스 다이어그램 이미지 개수가 충분히 크지 않지만, 도출된 실험 결과는 제안된 자동 분류 방식이 고려할 만한 가치가 있음을 나타낸다.

• 의학교육논단
• /
• 제17권1호
• /
• pp.33-38
• /
• 2015

During the education reform in 2009, the Catholic University of Korea College of Medicine (CUMC) adopted body systems as the basis for structuring basic medical education. After running the new program for 5 years, we need to evaluate the program by comparing it with nationwide standards. This study was designed to evaluate the coverage of our basic medical education program by comparing it with the assessment items of the medical licensing examination for physicians in the Republic of Korea. We built a relational database populated with 3,017 learning outcomes from all the courses on basic medical education. We tagged each learning outcome according to 2 criteria: 206 physician encounters and 9 outcome domains. A majority of the learning outcomes were in the domains of 'knowledge' and 'critical thinking'. In addition, we repeated the categorization process with 584 assessment items of the medical licensing examination in the Republic of Korea and compared them with the categorization results of the learning outcomes. Among the 206 physician encounters, we found that outcomes on family violence and sexual violence were missing in the learning outcomes of CUMC. Eighty-two physician encounters were associated with more than one outcome domain, and 96 physician encounters were covered in more than one course. Twenty-one physician encounters were repeated in 5 or more courses and 34 physician encounters had outcomes categorized into 3 or more domains. Thus, we showed that the 2-way categorization could be applied to the comparison and evaluation of two different education formats.

• 가정간호학회지
• /
• 제5권
• /
• pp.47-60
• /
• 1998

This study was done to identify basic information in classifying nursing diagnoses and nursing interventions needed for the further development of computerized nursing care plans. Data were collected by reviewing charts of 123 home care clients who had active disease, for whom at least one nursing diagnosis was on the chart, and who had been discharged. Data included demographics, medical orders, nursing diagnoses and nursing interventions. The results of the study, which found the most frequent medical diagnoses to be cancer (40.7%) and brain injury (26.8%), showed that 'Impaired Skin Integrity'(18.3%), 'Risk for Infection'(15.0%), 'Altered Nutrition, Less than Body Requirements'(13.8%), and 'Risk for Impaired Skin Integ rity'(9.9%) were the most frequent nursing diagnoses. 'Pressure Ulcer Care'(28.4%) was the most frequent intervention for 'Impaired Skin Integrity', 'Infection Protection'(16.0%) for 'Risk of Infection', 'Nutrition Counseling'(26.8%) for 'Altered Nutrition' and 'Positioning'(22.0%) for 'Risk for Skin Integrity Impairment', Comparison of interventions with the Nursing Intervention Classification(NIC) showed that the most frequent interventions were in the domain 'Basic Physiological' (33.94%), followed by 'Behavioral'(27.8%), and 'Complex Physiological' (22.6%). Interventions related to teaching family to give care at home could not be classified in the NIC scheme. Examination of the frequency of NIC interventions showed that for the domain 'Activity & Exercise Management', 75% of the interventions were used, but for seven domains, none were used. For the domain 'Immobility Management', 93% of the times that an intervention was used, it was 'Positioning', for the domain 'Tissue Perfusion Management', 'IV Therapy' (59.1%) and for the domain 'Elimination Management', 'Tube Care: Urinary'(54.0%). The nursing diagnoses 'Altered Urinary Elimination' and 'Im paired Physical Mobility' were both used with these clients, but neither 'Fluid Volume Deficit' nor 'Risk of Fluid Volume Deficit' were used rather 'IV Therapy' was an intervention for 'Altered Nutrition, Less than Body Requirements', A comparison of clients with cancer and those with brain injury showed that interventions for the nursing diagnosis 'Impaired Skin Integrity' were more frequent for the clients with cancer, interventions for 'Risk of Infection' were similar for the two groups but for clients with cancer there were more interventions for' Altered Nutrition'. Examination of the nursing diagnoses leading to the intervention 'Positioning' showed that for both groups, it was either 'Impaired Skin Integrity' or 'Risk for Skin Integrity Impairment'. This study identified a need for further refinement in the classification of nursing interventions to include those unique to home care and that for the purposes of computerization identification of the nursing activities to be included in each intervention needs to be done.

• 인터넷정보학회논문지
• /
• 제19권1호
• /
• pp.155-166
• /
• 2018

These days, online media, such as blogospheres, online communities, and social networking sites, provides the uncountable user-generated content (UGC) to discover market intelligence and business insight with. The business has been interested in consumers, and constantly requires the approach to identify consumers' opinions and competitive advantage in the competing market. Analyzing consumers' opinion about oneself and rivals can help decision makers to gain in-depth and fine-grained understanding on the human and social behavioral dynamics underlying the competition. In order to accomplish the comparison study for rival products and companies, we attempted to do competitive analysis using text mining with online UGC for two popular and competing ramens, a market leader and a market follower, in the Korean instant noodle market. Furthermore, to overcome the lack of the Korean sentiment lexicon, we developed the domain specific sentiment dictionary of Korean texts. We gathered 19,386 pieces of blogs and forum messages, developed the Korean sentiment dictionary, and defined the taxonomy for categorization. In the context of our study, we employed sentiment analysis to present consumers' opinion and statistical analysis to demonstrate the differences between the competitors. Our results show that the sentiment portrayed by the text mining clearly differentiate the two rival noodles and convincingly confirm that one is a market leader and the other is a follower. In this regard, we expect this comparison can help business decision makers to understand rich in-depth competitive intelligence hidden in the social media.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제18권1호
• /
• pp.89-104
• /
• 2024

Cloud-based e-learning (CBEL) represents a promising technological frontier. Existing literature has presented a diverse array of findings regarding the determinants that influence the adoption of CBEL. The primary objective of this study is to conduct an exhaustive examination of the available literature, aiming to determine the key predictors of CBEL utilization by employing the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology. A comprehensive review of 35 articles was undertaken, shedding light on the status of CBEL as an evolving field. Notably, there has been a discernible downturn in related research output during the COVID-19 pandemic, underscoring the temporal dynamics of this subject. It is noteworthy that a significant portion of this research has emanated from the Asian continent. Furthermore, the dominance of the technology acceptance model (TAM) in research frameworks is affirmed by our findings. Through a rigorous thematic analysis, our study identified five overarching themes, each encompassing a diverse range of sub-themes. These themes encompass 1) technological factors, 2) individual factors, 3) organizational factors, 4) environmental factors, and 5) security factors. This categorization provides a structured framework for understanding the multifaceted nature of CBEL adoption determinants. Our study serves as a compass, guiding future research endeavours in this domain. It underscores the imperative for further investigations utilizing diverse theoretical frameworks, contextual settings, research methodologies, and variables. This call for diversity and expansion in research efforts reflects the dynamic nature of CBEL and the evolving landscape of e-learning technologies.

• 디지털융복합연구
• /
• 제10권10호
• /
• pp.215-220
• /
• 2012

본 논문에서는 상황 변화 환경에서 적응적인 객체 인식을 위한 계층적 트리 구조를 이용한 조명 온톨로지 분류에 대한 방식을 제안한다. 본 논문에서는 상황이 불변하는 환경에서 동작하는 개발된 많은 시스템을 찾아냈고, 상황에 맞는 감지를 위한 새로운 개념의 트리 구조를 이용한 온톨로지를 도입하였다. 조명의 영향이 상황 인지 인식시스템을 아주 설계하기 어려운 시스템으로 만들기 때문에 본 논문에서는 트리 구조의 온톨로지를 사용하여 이러한 상황 변화 시스템을 설계하는데 더 중점을 두었다. 온톨로지는 일반적으로 사람들이 특정 분야의 것들에 대해 생각하는 방법의 추상적 모델에서 전형적으로 캡처된 한 분야의 개념화의 명시적 사양으로 정의 할 수 있다. 인간은 기본 원칙과 환경을 이해하고 설명하기 위해 온톨로지를 생성한다. 본 연구에서는 상황 온톨로지, 상황 모델링, 상황 적응 및 조명 기준에 따라 트리 구조 온톨로지를 설계하는 상황 분류를 제안했다. 조명 온톨로지의 적당한 영역을 선택한 후, 그 영역에서 더 나은 성능을 생산하는 동작의 한 집합을 선택하는데 있어서 장점을 얻었다. 본 논문에서는 역동적인 변화 환경에서 객체 인식의 영역에서 이러한 개념을 이용하여 폭 넓은 실험을 수행하였으며 제안하는 기본 개념에 대해 수행할 수 있는 많은 성공을 얻었다.