一种基于关联矩阵的时态关联规则挖掘算法

提出一种新的有别于将时间直接代入关联规则挖掘过程的方法。该方法分别针对确定的有效时间与不确定的有效时间两种情况,通过计算时态数据库中同一对象的元组中指定属性的有效时间之间的同时关系,将所有两两之间具有同时关系的归为一组,从而可以略去元组中的时间因素,并将时态数据库转化为事务数据库,然后即可利用传统挖掘算法进行挖掘。

An Associated Matrixes based Algorithm for Mining Temporal Association Rules

In contrast to the approaches directly introducing temporal elements into mining process, an associated matrixes based algo- rithm is proposed. By determining the simul relationship between valid times of the specified attribute in all the tuples that belong to an object, the tuples with simul relationship are combined to a same group. Therefore, temporal elements in the tuples can be omit- ted and a temporal database transferred to a transaction database, then the resulting databases mined by traditional algorithms.