Experiment about analyzing community graph which help us improve the influence maximization algorithm
Format the raw graph data. I provide three ways to generate the probability between two nodes.
./format -dataset path_to_dataset -flag your_flag
- flag = 0 probability is generated by 1/d(in), d(in) is the number of indegrees of a node'
- flag = 1 probability is generated by uniform distribution between [0, 1)
- flag = 2 probability is generated by normal distribution (mean, sigma*sigma)
Analyze the graph data, calculate the influence diffusion after serveral steps
./main -attr path_to_attrbute -graph graph
Attention: you can redirect the output stream to a file