会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 46. 发明公开
    • 도달 확률을 이용하여 링크 기반 유사도를 계산하는 방법 및 장치
    • 使用可达性计算基于链路的相似性的方法和装置
    • KR1020130103079A
    • 2013-09-23
    • KR1020120024421
    • 2012-03-09
    • 한양대학교 산학협력단
    • 김상욱김지수윤석호
    • G06F17/00G06F17/16
    • G06F17/15G06F17/16
    • PURPOSE: A method for calculating the link-based similarity by using reaching probability and a device thereof are provided to generate vectors corresponding to an object based on a link between the objects, to and use the cosine similarity between the vectors, thereby calculating the similarity between the objects. CONSTITUTION: A storage unit (410) stores information of objects, and a processing unit (420) obtains information of the objects from the storage unit and generates vectors corresponding to the objects. The processing unit selects two vectors of the vectors, calculates the cosine similarity between the selected vectors, and determines the similarity between the objects corresponding to the vectors based on the cosine similarity. A providing unit (430) provides the determined similarity, and the elements in the vectors correspond to one object of the objects. [Reference numerals] (410) Storage unit; (420) Processing unit; (430) Providing unit
    • 目的:提供一种通过使用到达概率来计算基于链路的相似度的方法及其装置,用于基于对象之间的链接生成与对象相对应的向量,并使用向量之间的余弦相似度,从而计算相似度 对象之间。 构成:存储单元(410)存储对象的信息,处理单元(420)从存储单元获取对象的信息,生成与对象对应的向量。 处理单元选择向量的两个向量,计算所选向量之间的余弦相似度,并且基于余弦相似度确定与向量相对应的对象之间的相似度。 提供单元(430)提供所确定的相似度,并且向量中的元素对应于对象的一个​​对象。 (附图标记)(410)存储单元; (420)处理单元; (430)提供单位
    • 48. 发明公开
    • 논문 랭킹 방법 및 장치
    • 排序纸的方法和装置
    • KR1020120088170A
    • 2012-08-08
    • KR1020110009325
    • 2011-01-31
    • 한양대학교 산학협력단
    • 김상욱배덕호황세미
    • G06F19/00
    • G06F17/30011G06F17/16
    • PURPOSE: A method and an apparatus for ranking paper are provided to calculate a rank of paper by reflecting issued year difference between papers in a reference relation. CONSTITUTION: A reference relation modeling unit(110) extracts a reference relation between papers. A rank calculating unit(120) calculates a rank of a first paper based on the reference relation. The rank calculating unit reflects difference between first paper issued year and second paper issued year. The rank calculating unit calculates the rank of the first paper. A core paper extracting unit(130) extracts the core paper based on each rank of one or more papers.
    • 目的:提供一种用于排名纸张的方法和装置,通过反映参考关系中的文件之间的年份差异来计算纸张的等级。 构成:参考关系建模单元(110)提取论文之间的参考关系。 等级计算单元(120)基于参考关系计算第一纸张的等级。 排名计算单位反映了第一份纸张发行年份和第二份纸张发行年度之间的差异。 秩计算单元计算第一纸的等级。 核心纸提取单元(130)基于一个或多个纸张的每个等级提取纸芯纸。
    • 50. 发明公开
    • 클러스터링과 특성분석을 이용한 구간 데이터에서 다차원 연관 규칙 마이닝 장치 및 방법
    • 采用聚类和表征的间隔数据采集多维度协会规则的装置和方法
    • KR1020100126988A
    • 2010-12-03
    • KR1020090045452
    • 2009-05-25
    • 한양대학교 산학협력단
    • 김상욱임승환권용석
    • G06F17/30G06F17/00
    • G06F17/30539G06F17/30592G06F17/30598G06F2216/03
    • PURPOSE: An apparatus and a method for mining of multi-dimensional association rules over interval data using clustering and characterization are provided to establish a section for drawing association rules having the highest reliability. CONSTITUTION: A sectioning unit(110) is sectioning a first attributes in the association rules. A clustering unit(120) hierarchically clusters the analysis object data based on the second attribute of the association rules. An association rule generating unit(130) uses the reliability of the association rules for each cluster, and processes an analysis of characteristic. In case the reliability of the association rules diminishes to the specific clustering, the association rule generating unit includes the association rules of the cluster in the result group of the association rules.
    • 目的:提供一种使用聚类和表征从间隔数据挖掘多维关联规则的装置和方法,以建立一个用于绘制具有最高可靠性的关联规则的部分。 规定:分割单元(110)正在关联规则中划分第一属性。 聚类单元(120)基于关联规则的第二属性对分析对象数据进行分层聚类。 关联规则生成单元(130)使用每个簇的关联规则的可靠性,并处理特征分析。 在关联规则的可靠性降低到特定聚类的情况下,关联规则生成单元包括关联规则的结果组中的聚类的关联规则。