1
کامپیوتر و شبکه::
نمره پرت
Different from the methods dis- cussed so far, it accepts a subgraph as an input query and returns top matching anomalous subgraphs from the original attributed network sorted by their outlier score.
Apart from learning the feature weights, lin- ear optimization is also used to calculate the outlier score of the subgraphs.
After matching the communities from successive snapshots, it generates outlierness score for every node in the community.
These outlierness scores are used to report the ECOutliers.
The algorithm that detects LEOutliers consists of two stages: (1) discovering Corenets based on the network topology and edge weights, and (2) measuring outlier score by inspecting and comparing Corenets at different snapshots.
واژگان شبکه مترجمین ایران