公开(公告)号：KR100679487B1

公开(公告)日：2007-02-07

申请号：KR1020060120267

申请日：2006-11-30

Applicant: 한국정보통신대학교 산학협력단

Inventor： 이관수 ,  선충현 ,  황태호 ,  임재형 ,  윤애영 ,  현순주 ,  노재정

IPC: G06F17/00 ,  C12Q1/68

公开(公告)号：KR100794516B1

公开(公告)日：2008-01-14

申请号：KR1020070124109

申请日：2007-12-03

Applicant: 한국정보통신대학교 산학협력단

Inventor： 이관수 ,  강지용 ,  윤태균 ,  선충현

IPC: G06F19/00

Abstract: A system and a method for diagnosis and clinical test selection using a case based machine learning inference are provided to offer information about accurate diagnosis by using an NNDT(Neural Network Decision Tree) teaching machine and a random forest teaching machine at the same time. A system for diagnosis and clinical test selection using a case based machine learning inference comprises a patient case database(200), an input device(100) and a machine learning classification device(400) constituted with one or more machine learning device(410). A machine learning trainer(300) determines important test items of each diseases, and stores it in a test items database(600). A diagnostor(500) including a test item selector(520) derives the main test item selection result of each diseases. An output device(700) outputs the test item selection result and a disease determination.

Abstract translation: 提供使用基于案例的机器学习推断进行诊断和临床测试选择的系统和方法，以通过使用NNDT（神经网络决策树）教学机和随机森林教学机同时提供关于准确诊断的信息。 一种使用基于病例的机器学习推断进行诊断和临床测试选择的系统包括由一个或多个机器学习装置（410）构成的患者病例数据库（200），输入装置（100）和机器学习分类装置（400） 。 机器学习训练器（300）确定每种疾病的重要测试项目，并将其存储在测试项目数据库（600）中。 包括测试项目选择器（520）的诊断器（500）导出每种疾病的主要测试项目选择结果。 输出设备（700）输出测试项目选择结果和疾病确定。

公开(公告)号：KR100806523B1

公开(公告)日：2008-02-21

申请号：KR1020070128834

申请日：2007-12-12

Applicant: 한국정보통신대학교 산학협력단

Inventor： 이관수 ,  김민성 ,  선충현 ,  한영웅

IPC: G06F9/50 ,  G06F15/16

CPC classification number: G06F9/5072 ,  G06F17/30734 ,  G06N5/04

Abstract: A semantic information-based grid management system supporting grid computing and a method thereof are provided to infer information for a program suitable for requested work and resource requirement, and generate a grid work detail file for assigning grid resources based on interred information by constructing application information in an ontology type, and forming the management system interposed between a user and grid middleware. An ontology data manager(140) defines ontology representing the information for an application field analysis and stores ontology data collected from an information provider to an ontology repository(130). A grid resource manager(180) collects and stores grid resource information of each grid resource to a grid resource table(190). Grid middleware(160) assigns a program to each grid resource according to grid work details and transfers a program execution result to the user. An inference engine(120) determines the optimal program for performing the analysis based on the ontology data and the grid resource information, and stores a grid resource list performing the optimal program. A grid work detail generator(150) makes the grid work detail by using the optimal program and the grid resource list.

Abstract translation: 提供了一种支持网格计算的基于语义信息的网格管理系统及其方法，以推断适用于所请求的工作和资源需求的程序的信息，并通过构建应用信息生成用于基于内部信息分配网格资源的网格工作细节文件 在本体类型中，并且形成插入在用户和网格中间件之间的管理系统。 本体数据管理器（140）定义表示用于应用程序字段分析的信息的本体，并将从信息提供者收集的本体数据存储到本体存储库（130）。 网格资源管理器（180）将每个网格资源的网格资源信息收集并存储到网格资源表（190）。 网格中间件（160）根据网格工作细节向每个网格资源分配程序，并将程序执行结果传送给用户。 推理引擎（120）基于本体数据和网格资源信息确定用于执行分析的最佳程序，并且存储执行最佳程序的网格资源列表。 网格工作细节生成器（150）通过使用最优程序和网格资源列表使网格工作细节。

公开(公告)号：KR100681795B1

公开(公告)日：2007-02-12

申请号：KR1020060119714

申请日：2006-11-30

Applicant: 한국정보통신대학교 산학협력단

Inventor： 이관수 ,  김민성 ,  선충현 ,  김진기 ,  김수영

IPC: G06F19/20 ,  G06F19/18 ,  G06F19/16

Abstract: A method for aligning a genome sequence in a grid computing environment and a program storing medium are provided to efficiently apply present sequence alignment programs difficult to compare a large quality of genome sequences owing to restriction of computing resources under the grid computing environment. The first and second genome sequences are cut by specific overlapped section based on calculating algorithm which aligns the size of the first and second sequences and the sequences, and repeated sequences in cut fragments are indicated(S100). The first and second genome sequence fragments are distributed to each computer of the grid computing environment, and are aligned by using the sequence alignment program(S200). Generated alignment results are added up and only the statistically meaning sequence information are extracted(S300).

Abstract translation: 提供了一种用于在网格计算环境中对齐基因组序列的方法和一种程序存储介质，以有效地应用由于网格计算环境下的计算资源的限制而难以比较大质量的基因组序列的当前序列比对程序。 基于计算算法对第一和第二基因组序列进行切割，所述计算算法将第一和第二序列和序列的大小对齐，并且指示切割片段中的重复序列（S100）。 将第一和第二基因组序列片段分发到网格计算环境的每个计算机，并通过使用序列比对程序进行比对（S200）。 生成的比对结果相加并且仅提取统计意义的序列信息（S300）。