摘要
论文提出了一种步长为2的宽度优先层次搜索的关联规则挖掘算法。该算法根据素数只有两个约数的性质,成功地把高维二进制事务库转化为一维事务库,把集合的运算转化为求约数的数值运算,把统计数据项频次的问题转化为统计事务的公约数。规则判断条件采用的是抽样平均可信度。同时,探讨了时间对关联规则挖掘产生的影响以及在导读系统中的应用。在仿真实验中,效果比较理想。
This paper puts forward an association rule mining algorithm of preferential level searching in the width of step 2 acording to the truth that prime number has only two divisors,the algorithm sucessfully changes high-dimension binary system routine assemble into one-dimension routine system,changes assemble operation,into numeral value operation of seeking divisors,changes counting data frequency into counting common divisors of routine sampling average credibility is employed in routine judging conditions.Meanwhile this paper discusses the influence of time on associations rule mining and the application in guide reading system.The experiment show the effects are more ideal.
出处
《计算机工程与应用》
CSCD
北大核心
2006年第33期191-194,共4页
Computer Engineering and Applications
关键词
关联规则
频集
素数存储
平均可信度
知识发现
Association rales
freguency itemset
prime namber storage
average credibility
knowledge discovery