在分析了当前基于距离的离群数据挖掘算法的基础上,提出了一种基于SOM的离群数据挖掘集成框架,其具有可扩展性、可预测性、交互性、适应性、简明性等特征。实验结果表明,基于SOM的离群数据挖掘是有效的。
Based on the analysis of the existing distance-based outlier detection algorithms, this paper proposed a SOM-based unifying framework for mining outliers, which had obvious superiority in scalability, predictability, interactiveness, adaptability, conciseness. Experimental results on real database show that the SOM-based outlier mining is effective.