首页> 外国专利> CLUSTER ANALYZING AND PROCESSING METHOD AND RECORDING MEDIUM HAVING CLUSTER ANALYZING PROGRAM RECORDED THEREON

CLUSTER ANALYZING AND PROCESSING METHOD AND RECORDING MEDIUM HAVING CLUSTER ANALYZING PROGRAM RECORDED THEREON

机译:聚类分析和处理方法及记录介质,并在其上记录有聚类分析程序

摘要

PROBLEM TO BE SOLVED: To set the order of the number of calculating times to N (number of data) at the minimum at the time of performing cluster analysis and processing. ;SOLUTION: The matrix {Aij} (i, j=1 to N) of the degree of resemblance between the data of a set D composed of N-pieces of data is calculated. The maximum degree of resemblance and the row ID of the matrix {Aij} at that time are found by successively retrieving the row IDs of the matrix {Aij} from the top and stored in a next candidate storing table T. Then the maximum value of the matrix {Aij} is found by successively retrieving the table T from the top and a new cluster is generated and added to the matrix {Aij} and the table T. Thereafter, old data (the data of the row ID and column ID containing the highest degree of resemblance) are deleted from the matrix {Aij} and table T. The above-mentioned operations are repeated.;COPYRIGHT: (C)2001,JPO
机译:解决的问题:在执行聚类分析和处理时,将计算次数的顺序设置为最少N(数据数)。 ;解决方案:计算由N个数据组成的集合D的数据之间的相似度的矩阵{Aij}(i,j = 1至N)。通过从顶部连续检索矩阵{Aij}的行ID并将其存储在下一个候选存储表T中,可以找到此时最大相似度和矩阵{Aij}的行ID。通过从顶部连续检索表T来找到矩阵{Aij},并生成一个新簇并将其添加到矩阵{Aij}和表T中。此后,旧数据(包含行ID和列ID的数据最高相似度)从矩阵{Aij}和表T中删除。重复上述操作。;版权:(C)2001,JPO

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号