公开(公告)号：KR1020120136677A

公开(公告)日：2012-12-20

申请号：KR1020110055750

申请日：2011-06-09

Applicant: 성균관대학교산학협력단

Inventor： 김응모 ,  김영희

IPC: G06F17/30 ,  G06F17/10

CPC classification number: G06F17/30516 ,  G06F17/10 ,  G06F17/30327

Abstract: PURPOSE: A data stream frequent occurrence pattern extraction method based on a weighted value and data base tree structure thereof are provided to offer meaningful information by extracting a frequent occurrence pattern based on a weighted value. CONSTITUTION: A data stream transaction including a window composed of articles is scanned and is calculated by obtaining a weighted value of the articles(S1). A weighted value frequent occurrence article is extracted by creating a frequent pattern database tree structure based on the weighted value which is determined through a descending order in windows(S6). The frequent occurrence pattern is extracted by scanning the articles including the weighted value. [Reference numerals] (AA) Start; (BB) End; (S1) Calculating data stream scan and weighted value support degree of each article; (S2) User definition minimum weighted value support degree threshold value and minimum weighted value support error threshold value definition step; (S3) Window size and weighted value range definition, maximum in article, minimum weighted value support calculation step; (S4) Weighted value support frequency pattern classification step; (S5) Weighted value support frequency pattern tree creation step; (S6) Weighted value support frequency article extraction step; (S7) Data stream update step

Abstract translation: 目的：提供基于加权值和数据库树结构的数据流频繁发生模式提取方法，通过基于加权值提取频繁发生模式来提供有意义的信息。 构成：扫描包括由物品组成的窗口的数据流交易，并通过获得物品的加权值来计算（S1）。 通过基于通过窗口中的降序确定的加权值创建频繁模式数据库树结构来提取加权值频繁发生物品（S6）。 通过扫描包括加权值的物品来提取频繁发生模式。 （附图标记）（AA）开始; （BB）结束; （S1）计算每篇文章的数据流扫描和加权值支持度; （S2）用户定义最小加权值支持度阈值和最小加权值支持误差阈值定义步骤; （S3）窗口大小和加权值范围定义，文章最大值，最小加权值支持计算步骤; （S4）加权值支持频率模式分类步骤; （S5）加权值支持频率模式树创建步骤; （S6）加权值支持频率文章提取步骤; （S7）数据流更新步骤