目前的数据挖掘基本上都是基于普通数据集的挖掘,针对星型模式结构的数据挖掘的研究工作较少,为此定义星型模式挖掘结构,并在此基础上构建一种关联规则挖掘算法,该算法先扫描事实表,产生最大频繁项集和关联规则,进而以此为基础,提出一种基于连接条件和关联规则局部有效性的理论,并在此基础上建立一种快速扫描维表属性的方法,一次产生维表隐藏的关联规则,这个扫描是基于局部的,不是基于全局的,同时可根据需要,对于不明确的关联规则,通过构建扩展的维表,进行隐知识的挖掘.算法挖掘速度快,若合理地构建扩展维表,能够发现扩展的隐藏信息.%Current data mining is based on the mining of general data set basically. The research to data mining of the star schema structure is less. So the star schema mining structure is defined, and based on which an mining algorithm with association rules is constructed. The algorithm first scans the fact table, and produces maximal frequency item sets and association rules, with which as the basis, the theory based on the local efficiency principle of linking conditions and association rules is put forward, and the method scans the dimension table attributes quickly. It produces the association rules one-off. The scan is based on the part, not global. At the same time the undefined association rules are dealt with by mining the implicit knowledge through constructing extended dimension table. The mining speed of the algorithm is faster. Through building expanded dimension table reasonably, the extended hidden information can be found in this way.
展开▼
