一种加权时态关联规则挖掘算法

Mining algorithm of weighted temporal association rules

许多现实数据库都存在时态语义问题,因此在挖掘关联规则时附加上时态约束会使规则更具有实际意义。但目前提出的大多数时态关联规则挖掘算法,一般都认为每个数据项的重要性相同,而从决策者角度出发,往往会优先考虑利润较高的项目。提出了一种加权时态关联规则挖掘算法,以项目的生命周期作为时间特征,允许用户设定不同的项目权重。实验结果证明,该算法不仅能有效地发现加权时态关联规则,而且挖掘出的规则更有价值。

The association rules would be more real and useful when considering temporal constraint because of temporal semantics existing in many real databases.Lots of algorithms for mining temporal association rules have been proposed at present and most of them treat each item uniformly,but decision-makers are more inclined to items whose profits are higher than others.A mining algorithm of weighted temporal association rules is proposed.In the algorithm item＇s lifespan is used as time feature and user is allowed to specify different weights for different items.Experiments show that not only the new algorithm is effective in mining the weighted temporal association rules but also more useful association rules can be found.