This and java implementation of the method for finding overlapping communities in network. Below you can find an example of method output for "School friendship" dataset.
- ./Scripts – directory for all Jupyter notebook that were used for computation experimenst
- ./final_pictures – contains all pictures and gephi files.
- ./lpam – source code of lpam method
- ./datasets – contains realworld datasets and sysntetic datasets
- ./related_methids –
- ./literature – papers about overllaping communities detection methods in networks
In order to repoduce all computation experiments you should compile all method in the related_methods directory and tool for measuring ONMI in the directory Overlapping-NMI. Just go to corresponding folder and type "cmake".
cd Overlapping-NMI
make
To find exact solution for k-medois the installation of lp_solver is required. ...TODO
- Leonidas Pitsoulis ([email protected]): idea of overlapping communities detection based on the link partitoing with the help of non-overlapping communities detection methods based on the partitioning around medoids
- Marat Shamshetdinov ([email protected]): Implementation of the exact model for P-Median prolem for lp_solver and cplex solvers.
- Nikit Putehin: implementation of amplified commute distance and heuristics method for findg k-medoids (Clarance and k-meanns heuristics)
- Alexander Ponomarenko ([email protected]): Basic implementation of link partitionig algorithm on java, calculating commute distance, jupyter notebooks and collecting everything together
