This repository contains a reference implementation of the algorithms for the paper:
Longlong Lin, Pingpeng Yuan, Rong-Hua Li, Jifei Wang, Ling Liu, Hai jin. Mining Stable Quasi-Cliques on Temporal Networks. IEEE Transactions on Systems Man Cybernetics-Systems
Codes run on C++11 or later. When we run codes on linux, -O2 is recommended to optimizate compilation.
-
The input file should be started with V and E
-
V: upper limit of node id or higher
-
E: the number of edges
-
-
For each event, there are 3 parameters: vertex1, vertex2, time
-
vertex1, vertex2: the vertices of the changing edge
-
time: the time when the event happens
-
-
the time of any (vertex1,vertex2) must be sorted in ascending order
-
There are two output files
-
TGRA.txt : store the spent time of TGRA algorithm
-
res.(dataset) : dataset refers to dataset name, stroe the spent time of B&B algorithm
-
-
No matter how many times BB.cpp is excuted, we can get only one TGRA.txt file. That means TGRA.txt stores all TGRA algorithm result.
-
res.(dataset) is different, results of different parameters of one dataset share one dataset. That means one dataset outputs one res.(dataset) file.
-
create an excutable application by compiling "BB.cpp", for example:
g++ -o test BB.cpp
-
run the excutable application with 5 parameters, which are input file name, gamma, mode, rho, theta, for example:
./test enron 0.8 4 0.6 3
-
input file name: the name of the file describing the tempral graph
-
gamma
-
mode: five pruning algorithms
-
tho
-
theta
-