Skip to content

Dynamic Social Network Simulation Data with Ground Truth Labels and Features

License

Notifications You must be signed in to change notification settings

AkandaAshraf/VirtualSoc

Repository files navigation

VirtualSoc

Simulate Dynamic Social Networks with ground truth labels and features.

Required packages:

numpy, pandas, SciPy, SALib, pathos, PyPrind

cupy and CUDA for the GPU version.

For a single network simulation follow the ScriptSingleNetwork.py
For multiple networks simulation follow the ScriptMultiNetwork.py

If you have an NVIDIA GPU with CUDA, use the GPU branch, otherwise the CPU branch

Thanks and happy simulation.

The paper (preprint) : https://arxiv.org/abs/1905.09087 (Simulation and Augmentation of Social Networks for Building Deep Learning Models)

p.s. to calculate the generated network's properties and statistics: To calculate network statistics and properties for the generated networks you can use the R script. Use the function pipeNetworkStats("D:/VirtualSocPP1/", threads=7) , and pass the root directory path to the function and number of threads you want it to use. There are dependencies for the r script and they need to be installed to run the script. (This R script is separate from this project and relies heavily on other libraries for graph properties algorithm. The R script is provided for calculating graph properties but not a part of the VirtualSoc project)

Few generated sample datasets are uploaded to the repo. data_sample.zip

Many more generated datasets from VirtualSoc https://www.kaggle.com/akandaashraf/virtualsoc1