中文摘要：

提出一种具有通用性的分布式异常检测框架.首先,利用本地的无监督异常检测算法,建立多个本地检测模型;然后,将各个本地无监督检测模型转换成统一的共享模型;最后,采用集成学习的方法,综合考虑各模型差异性和准确性,实现全局异常检测.实验结果表明,基于模型共享的分布式异常检测不仅能有效地保护数据隐私,减少通信开销,同时能获得和集中式检测相当甚至在某些情况下更好的效果.

英文摘要：

A novel general framework for distributed anomaly detection is proposed. Local models are built for distributed data sources with unsupervised anomaly detection methods. Local models are transformed into uniform models, and then learned models are reused for new data and combining their results by considering both quality and diversity of them to detect anomalies in a global view. The results show that, the proposed distributed anomaly detection method achieved prediction performance comparable to, or slightly better than the global anomaly detection algorithm applied on the data set obtained when all distributed data sets are merged.