A Dynamic Community Detection Method for Complex Networks Based on Deep Self-Coding Network
Algorithm 1
Community detection algorithm.
INPUT: Graph corresponding to continuous timestamp ;
OUTPUT: Community structure under timestamp ;
(1)
By extracting a variety of information from the dynamic graph , the information is fused into a time matrix ;
(2)
;
(3)
Set time matrix , is the number of neural network layers and is the number of iterations;
(4)
set ;
(5)
for do ;
(6)
for do ;
(7)
The stack self-coding network is constructed, and the encoder in the stack self-coding network is used to extract the feature of the graph information;
(8)
Obtain potential characterization information
(9)
The parameters , are optimized according to iterative update formula (4);
(10)
set ;
(11)
End for
(12)
End for
(13)
Decode the extracted potential representation and output the predicted community structure ;
(14)
Cluster and output the community detection results under the current network structure;
