1
کامپیوتر و شبکه::
ماتریس باقی مانده
The authors propose a non-negative residual matrix factorization framework (NrMF) to identify anomalous nodes and edges.
are factors and R is the residual matrix.
The low-rank factors F and G show the community structure in the network, whereas the residual matrix R often indicates anomalies in networks.
Unlike the traditional non-negative matrix factorization (NMF) that re- stricts the factors F and G to be non-negative, NrMF enforces non- negativity constraint on the residual matrix R for finding anomalies.
(2015) proposed a framework based on the principal eigenspace of the network's residuals matrix in which an observed random network is compared to its expected value to find anomalous subgraphs.
واژگان شبکه مترجمین ایران