发明名称 INFORMATION RETRIEVAL PROGRAM, INFORMATION RETRIEVAL DEVICE AND INFORMATION RETRIEVAL METHOD
摘要 PROBLEM TO BE SOLVED: To improve a retrieval speed by narrowing down clusters having possibility of matching an inquiry graph. SOLUTION: Clustering means to divide the whole retrieval target graph into a plurality of clusters (subgraphs) as shown in the following definitions (1-1)-(1-4). (1-1) A node is included in (only) one of the clusters. (1-2) The cluster is a linked graph. (1-3) The number of the links straddling the clusters is reduced as thoroughly as possible. (1-4) The whole is divided with good balance. That is, the division is performed such that the number of the nodes included in each cluster becomes as equal as possible. By the clustering, the retrieval target graph 100 is divided into the clusters C<SB>1</SB>to C<SB>3</SB>. The clustering clearly satisfies the definitions (1-1) to (1-4) that are properties desired to the clustering. COPYRIGHT: (C)2010,JPO&INPIT
申请公布号 JP2009258794(A) 申请公布日期 2009.11.05
申请号 JP20080103835 申请日期 2008.04.11
申请人 FUJITSU LTD 发明人 YAMANE YASUO;IGATA NOBUYUKI
分类号 G06F17/30 主分类号 G06F17/30
代理机构 代理人
主权项
地址