Pellet本体概念分类算法优化

Optimization of the Algorithm for Building Ontology Concept Classifications in Pellet

针对Pellet1.4(简称Pellet)推理机的本体概念分类算法效率较低,为了通过减少调用ABOX推理概念间包含关系来提高分类性能,提出了一种全局标记方法。方法首先对所有节点嵌入一个二维散列mark,用于节点存储与其他节点的包含关系标记,然后在推理过程中通过ABOX的推理结果,动态修改相关节点中mark的标记信息。最后充分利用以前的推理结果所修改的标记信息来减少ABOX一致性检测次数。通过使用全局标记方法的算法替代Pellet推理机中原有概念分类算法对实际本体分类进行仿真,实验结果表明该方法可以有效地减少调用ABOX执行一致性检测次数,提高本体概念分类性能。

In this paper, the ontology concept classification algorithm of Pellet is researched first, but this algorithm affecting the performance of the two concepts＇ subsumption by ABOX which will drawback the performance of algorithm. For improving the performance of algorithm, an optimized algorithm with respect to classification is proposed based on global marking. The experimental results indicate that the optimization can effectively decrease the number of consistent checking in Abox and improve the performance of classification.