摘要
复杂数据集通常包含不同的组织模式,传统的离群检测算法从单一视角寻找离群点,不能充分利用多视角信息,造成信息遗漏。提出一种基于多视角聚类的离群检测算法,该算法一方面采用谱聚类,以确保高质量的聚类结果;另一方面通过希尔伯特-施密特独立性准则,以确保新的聚类结果相对于已知划分模式是无冗余的。对得到多个视角进行离群分析,从而得到更准确的离群集。研究结果表明,该算法能够提高离群检测精度。
Complex data sets usually contain different organization patterns,traditional outlier detection algorithm cannot make full use of multi-view information and cause outlier missing from a single point of view.Proposes a multi-view clustering outlier detection algorithm,which on the one hand uses a spectral clustering algorithm to ensure high quality of clustering results;on the other hand,through Hilbert-Schmidt independence criterion to ensure that the new clustering result and the known partition model comparison are not redundant.And then gets more accurate outlier sets through the multi-view of the outlier analysis.The results show that the algorithm can improve the accuracy of outlier detection.
关键词
离群检测
多视角
谱聚类
希尔伯特-施密特独立性准则
Outlier Detection
Multi-View
Spectral Clustering
Hilbert-Schmidt Independence Criterion
