• 정보처리학회논문지D
• /
• 제9D권2호
• /
• pp.215-226
• /
• 2002

미디어의 발달과 생활 패턴의 변화를 토대로 새롭게 나타나고 있는 다양한 판매 패턴들을 분석하는데 있어 단일한 분석 방법을 적용하는 것은 효과적이지 못하다. 특히 신선 식품이나 기념일 주변에서 집중적인 매출이 발생하는 품목들은 제한된 시간 내에 판매를 최대로 해야 하는 시간적 제약을 갖는다. 그러나 기존의 연관규칙 탐사 기법은 대규모 거래 데이터베이스로부터 반복적 스캔 연산을 통해 연관규칙 탐사를 수행하기 때문에 제한된 시간안에서 빈번히 필요로 하는 패턴을 분석하기에는 비효율적이기 때문이다. 따라서 이 논문에서는 시간 제약을 갖는 특수한 판매 패턴에 대한 실시간 연관규칙 탐사가 가능하도록 하기 위해 트리거와 저장 프로시져를 이용한 점진적 후보항목 관리 모델을 제안한다. 아울러 이 논문에서는 제안 모델의 구현 및 실험을 통해 그 성능 특성의 분석도 수행한다. 특히 이 논문에서 제안하는 방법은 이중 해쉬 기법을 이용함으로써 연산의 성능을 향상시킨다.

• Journal of Information Processing Systems
• /
• 제8권4호
• /
• pp.575-588
• /
• 2012

Exploiting unlabeled text data with a relatively small labeled corpus has been an active and challenging research topic in text mining, due to the recent growth of the amount of biomedical literature. Biomedical named-entity recognition is an essential prerequisite task before effective text mining of biomedical literature can begin. This paper proposes an Active Co-Training (ACT) algorithm for biomedical named-entity recognition. ACT is a semi-supervised learning method in which two classifiers based on two different feature sets iteratively learn from informative examples that have been queried from the unlabeled data. We design a new classification problem to measure the informativeness of an example in unlabeled data. In this classification problem, the examples are classified based on a joint view of a feature set to be informative/non-informative to both classifiers. To form the training data for the classification problem, we adopt a query-by-committee method. Therefore, in the ACT, both classifiers are considered to be one committee, which is used on the labeled data to give the informativeness label to each example. The ACT method outperforms the traditional co-training algorithm in terms of f-measure as well as the number of training iterations performed to build a good classification model. The proposed method tends to efficiently exploit a large amount of unlabeled data by selecting a small number of examples having not only useful information but also a comprehensive pattern.

• 한방안이비인후피부과학회지
• /
• 제36권1호
• /
• pp.50-63
• /
• 2023

Objectives : This study used a network pharmacology approach to explore the active compounds and therapeutic mechanisms of Yijin-tang on Meniere's disease. Methods : The active compounds of Yijin-tang were screened via the TCMSP database and their target proteins were screened via the STITCH database. The GeneCard was used to establish the Meniere's disease-related genes. The intersection targets were obtained through Venny 2.1.0. The related protein interaction network was constructed with the STRING database, and topology analysis was performed through CytoNCA. GO biological function analysis and KEGG enrichment analysis for core targets were performed through the ClueGO. Results : Network analysis identified 126 compounds in five herbal medicines of Yijin-tang. Among them, 15 compounds(naringenin, beta-sitosterol, stigmasterol, baicalein, baicalin, calycosin, dihydrocapsaicin, formononetin, glabridin, isorhamnetin, kaempferol, mairin, quercetin, sitosterol, nobiletin) were the key chemicals. The target proteins were 119, and 7 proteins(TNF, CASP9, PARP1, CCL2, CFTR, NOS2, NOS1) were linked to Meniere's disease-related genes. Core genes in this network were TNF, CASP9, and NOS2. GO/KEGG pathway analysis results indicate that these targets are primarily involved in regulating biological processes, such as excitotoxicity, oxidative stress, and apoptosis. Conclusion : Pharmacological network analysis can help to explain the applicability of Yijin-tang on Meniere's disease.

• 동의생리병리학회지
• /
• 제34권4호
• /
• pp.184-190
• /
• 2020

This study was performed to find antibacterial substances contained in Scutellariae Radix (SR) using a systems pharmacological analysis method and to establish an effective strategy for the prevention and treatment of infectious diseases. Analysis of the main active ingredients of SR was performed using Traditional Chinese Medicine Systems Pharmacology (TCMSP) Database and Analysis Platform. 36 active compounds were screened by the parameter values of Drug-Likeness (DL), Oral Bioavailability (OB), and Caco-2 permeability (Caco-2), which were based on the drug absorption, distribution, metabolism, and excretion indicators. The UniProt database was used to obtain information on 159 genes associated with active compounds. The main active compounds with antibacterial effects were wogonin, β-sitosterol, baicalein, acacetin and oroxylin-A. Target proteins associated with the antibacterial action were chemokine ligand 2, interleukin-6, tumor necrosis factor, caspase-8,9 and mitogen-activated protein kinase 14. In the future, systems pharmacological analysis of traditional medicine will be able to make it easy to find the important mechanism of action of active substances present in natural medicines and to optimize the efficacy of medicinal effects for combinations of major ingredients to help treat certain diseases.

• 대한한의학방제학회지
• /
• 제18권2호
• /
• pp.227-239
• /
• 2010

Objective : The purpose of this study was to develop the simulator which can analyze the anti-inflammatory effects of herbs based on e-cell, or the virtual cell. Method : We have ensured the medical herbs and its active compounds by investigating the oriental medicine records and NBCI(Biomedicine database). Also we have developed the web-based search system for confirming database related to anti-inflammation. We have researched the cell signal pathway related with inflammatory response control and established the mathematical model of herb interaction on selected signal pathway in e-cell. Finally we have developed the prototype which can confirm the result of this model visibly. Results : We constructed the database of 62 cases of anti-inflammatory active compounds in 61 cases of medical herbs which have been known anti-inflammation effects in the paper, 16 cases of inflammatory factors, 10 cases of signal pathways related with inflammatory response and 6,834 cases of URL(Uniform Resource Locator) of referenced papers. And we embodied the web-based research system, which can research this database. User can search basic and detailed information of medical plants related with anti-inflammatory by using information system. And user can acquire information on an active compounds, a signal pathway and a link URL of related paper. Among investigated ten pathways, we selected NF-${\kappa}B$, which plays important role in activation of immune system, and we searched the mechanisms of actions of proteins which could be components of this pathway. We reduced total network into IKK-$I{\kappa}B$ - NF-${\kappa}B$, and completed mathematic modeling by using ordinary differential equations and response variables of $I{\kappa}B-NF-{\kappa}B$ signaling model network which is suggested by Baltimore Group. We designed OED(Ordinary Differential Equation) for response of IKK, $I{\kappa}B$, $NF-{\kappa}B$ in e-cell's cytoplasm and nucleus, and measured whether an active compound of medicinal plants which is inputted by an user would have a anti-inflammation effects in obedience to change in concentration over time. The proposed model was verified by using experimental results of the papers which are listed on NCBI.

• 한국원자력학회:학술대회논문집
• /
• 한국원자력학회 1996년도 춘계학술발표회논문집(1)
• /
• pp.374-379
• /
• 1996

경보 처리 시스템 (Alarm Processing System)은 발전소의 이상 상태를 검출하여 운전원에게 알려 적절한 조치를 취하게 함으로써 발전소의 상태를 안전하게 유지하도록 하는 운전지원시스템이다. 본 논문은 이러한 목적을 보다 효과적으로 달성하기 위하여 경보 처리 시스템을 경보 생성과 경보 분류, 경보 표시의 3 단계로 설계하였으며, 이를 기반으로 능동 데이터베이스 (Active Database)의 능동 기능을 이용한 개선된 경보 처리 시스템 (Advanced Active Alarm Processing System)을 개발하였다. AAAPS는 데이터 입력의 변경에 대하여 모듈화된 경보 처리 프로그램을 능동적으로 수행시키며 생성된 경보는 윈도우를 이용한 개선된 MMI (Man Machine Interface)로 운전원에게 제공한다.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1997년도 추계학술대회 논문집
• /
• pp.437-441
• /
• 1997

In the conventional remote monitoring system, a user in front of a computer monitor can acquire only 2 dimensional visual information in a passive way. Thus, even thoght the user finds an interesting object from the video image, helshe can hardly acquire additional information on the object such as name. 311 shape, etc. In this paper, an active monitoring system that shows additional information on the selected object is proposed. The active remote monitoring system can calculate the 3D position of the object that is selected in the video images. Then, using the 3D position of the object, other information on the object can be retrieved from the database and shown on the screen. To calculate the 3D position of the object, 2 CCD cameras that can be tilted and panned using 3 stepping motors are used. The algorithm of 3D position calculation and the result of experiments are explained.

• 정보처리학회논문지A
• /
• 제18A권4호
• /
• pp.159-164
• /
• 2011

본 논문에서는 다양한 센서 네트워크 응용을 위한 능동적 에이전트 미들웨어를 제안한다. 이를 위하여, 이동 에이전트가 네이밍 에이전트의 네임 스페이스의 메타 테이블로부터 제공되는 이주 대상 센서 노드들을 차례로 방문하여, 사용자 조건, 용도 및 필요에 따른 다양한 능동규칙을 통하여 센서 데이터를 수집 및 전송하며, 임의의 발생 사건들(센서 데이터 값 및 시간 등의 변화)에 상응하는 조치들을 직접 실행할 수 있는 능동규칙 탑재 이동 에이전트를 설계 구현한다. 또한, 기존의 센서 데이터 관련 규칙 및 상황 데이터베이스 시스템과의 연동을 통하여 다양한 능동적 센서 네트워크 응용에의 적용 가능성을 보인다.

• 전자공학회논문지CI
• /
• 제38권3호
• /
• pp.41-54
• /
• 2001

본 논문은 시간 데이터베이스에 능동 규칙 시스템을 결합한 능동적 시간 데이터베이스의 능동 규칙 및 실행 모델을 제안한다. 기존의 시간 데이터베이스 능동 규칙 연구들이 시간 릴레이션에 대한 조건-조치 형태의 규칙 지원에 머물렀던데 비하여 제안하는 능동 규칙 모델은 유효 시간 및 트랜잭션 시간을 지원하는 이원 시간 릴레이션에 대해 SQL3 트리거의 기능 모두를 시간 데이터베이스 상에서 지원한다. 즉, 전이 릴레이션 및 전이 변수, 조건-조치의 튜플 및 문 단위 수행을 포함한다. 또한, 능동 규칙의 유효 시간 정의 및 유효시간 사건 등의 새로운 개념을 제시한다.

• 대한안전경영과학회지
• /
• 제12권3호
• /
• pp.27-35
• /
• 2010

In this study, we have developed the database of safety and health information for pesticide active ingredients used in Korea. There were 1,283 pesticide items among which 296 were found to be out of use in current. A total of 349 pesticide ingredients were being used in Korea. The database consists of 32 types of information including chemical characteristics, acute toxicity, chronic toxicity (carcinogenic and reproductive toxicity), specific symptoms by exposure route and first aid. When pesticide ingredients were assessed in terms of key properties such as color, odor, acute toxicity, carcinogenic and reproductive toxicity, they were white, colorless and odorless, in general. When ingredients were classified by category of acute toxicity, 'Non-hazardous' represented 29%, followed by 'Slightly hazardous' at 16%, 'Moderately hazardous' at 14%, 'Highly hazardous' at 5%, and 'Extremely hazardous' at 2%. 85 out of 349, or 24% of ingredients were found to be possibly carcinogenic to human. This database is expected to provide an easy access for farmers, agriculture supervisors, researchers and consumers, and it can ultimately be used as basic data on farmer's safety and health.