基于图的异常检测

1. Graph mining

Among various established graph mining tasks, the significant ones are :

1. Subgraph Discovery

• frequent subgraph discovery(FSD)
• dense subgraphs discovery(DSD),

2. graph pattern matching,

3. community/group detection.

2. Analyzing Graph/Network Data

Neighborhood-Based Analysis

在基于邻居的图研究方法中，有一个比较重要的概念，叫“egonet(自我中心网络)”。

Community Detection for Analysis

3.Types of Anomalies

3.1 Anomalous Vertices

3.2 Anomalous Edes

3.3 Anomalous Subgraphs

3.4 Event and Change Detection