Implementation for Reliable Community Search

This repository is a reference implementation of the querying algorithms proposed in "Reliable Community Search in Dynamic Networks".

Data Source

Popular Temporal/Dynamic Graph Data are Available at:

• Stanford Large Network Dataset Collection
• Network Data Repository

Data preparation

Data is organized in .gml format where each file represents a one-timestamp graph instance and each edge is attached with a "weight" attribute.

Requirements

• Python 3
• networkx == 2.6.3
• click == 8.0.3

Run the Code

Input the dataset name, parameters and choose the algorithm to run the code

python run.py

Parameters

• Dataset name: name of the dataset folder, string
• $\theta$ (Theta): parameter of the edge weight threshold, float number in [0,1].
• $k$ (K): parameter of the k-core constraint, integer
• $q$ (Query): query vertex, string format
• $\alpha$ (Alpha): parameter to balance the importance of community size and duration, positive float
• $T_s$ (T_s): starting timestamp of the query interval (included), integer
• $T_e$ (T_e): ending timestamp of the query interval (excluded), integer

Input:

Dataset name(str): Bitcoin_otc
Theta(float): 0.1
K(int): 2
Query(str): 1
Alpha(float): 5
T_s(int): 0
T_e(int): 10
Type one number to chose the algorithm (int): [1]EEF; [2]WCF;: 2

Output:

Index construction time: 3.3389127254486084
Running time of WCF query: 0.4596281051635742
CRC identified with size 16 and time interval [3,4]
CRC output at:  Bitcoin_otc.output-0.1-2-1-5.0_WCF

Visualization Example

demo of (0.5, 4)-core community querying "funny" on dataset Reddit

blue dash lines represent the edges whose weights are less than 0.5, and black lines represent the edges whose weights are no less than 0.5.