摘要
提出了一种基于关联规则二进制数组密集树的数据挖掘算法——BSC-tree算法,该算法将传统的交易数据库简化成二进制序列;其次在此基础上构造出BSC-trees,作为数据挖掘的准备条件,并由此求出BSC-trees的所有路径码,进而通过BSC-trees路径码间的逻辑与运算,他一次只需扫描整个数据库就能够快速地求出所有的关联规则。并将该算法用于交通事故历史数据的挖掘,他能够提取隐含在交通事故中未知的有用信息,为分析交通事故中各种诱因提供辅助性的决策。试验结果表明该算法优于其他算法。
The text proposes a bit string compression tree(BSC- tree) algorithm based on association rules which firstly simplies the traditional database into a bit string;secondly constructs BSC- trees as"data mining ready". Therefore produces the BSC- trees'path codes,which can produce all the frequent itemsets through scaning the database only once. Applying the algorithm to mine the historical datas of traffic accident which can bring out the unknown useful information hided in the traffic accidents to provide decision - making for analying the causes of all kinds of traffic accidents. The experimental results show that the BSC - trees algorithm is better than other algorithm.
出处
《现代电子技术》
2007年第20期71-74,共4页
Modern Electronics Technique
基金
国家科技部"十五"攻关项目(NO.2001DA204B01-03)
关键词
数据挖掘
关联规则
二进制数组密集树
交通事故
data mining
association rules
bit string compression tree
traffic accident