Skip to content

"Fast Computation of Dense Temporal Subgraphs."

Notifications You must be signed in to change notification settings

hurenjun/TDSubgraph

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

View under MarkDown mode for better representation.

Basic Information

  • Project: TDSubgraph, i.e., FIDES in paper
  • dev: Renjun Hu
  • homepage: https://hurenjun.github.io/
  • Reference: "Fast Computation of Dense Temporal Subgraphs." In ICDE, 2017.
  • 2017/1/17: reliease the first version of TDSubgraph
  • 2017/6/14: fix the bug in IntervalSelector that it throws an out-of-bounds exception when the temporal graph contains only single snapthot. The bug is reported by Yanyan Zhang from SFU.
  • 2017/10/16: release the beta version if FIDES in subdirectory 'FIDES_plus'.

Usage

  • run: compute dense subgraph of temporal graph

    • E.g.: -run traffic-small.txt 0 10 10 1.0
    • traffic-small.txt: temporal graph
    • 0 & 10: beginning & ending timestamps of temporal graph (from 0 to 10 (including))
    • 10: number k of time intervals that FIDES considers
    • xi: smoothing factor (pg 4 in paper), 1.0 by default
  • runIntv: compute dense subgraph of temporal subgraphs given time intervals (interval.txt)

    • E.g.: -runIntv traffic-small.txt 0 10 10 1.0
    • Parameter description is the same to run
    • 0 & 10 are the beginning and ending timestamps of the temporal graph that we load into main memory.
    • It will compute dense subgraph of temporal subgraphs whose beginning and ending timestamps are given in interval.txt.
  • computeADS: compute dense subgraph on aggregate graph

    • E.g.: -computeADS traffic-small.txt 0 10
    • Parameter description is the same to run
    • It will compute dense subgraph of aggregate graphs of temporal subgraphs whose beginning and ending timestamps are given in interval.txt
  • ADSOPT: compute dense subgraph on aggregate graph, with 4 combinations of optimizations

    • E.g.: -ADSOPT traffic-small.txt 0 10
    • Parameter description is the same to run
  • ecrate: compute ec rate (i.e., p_{EC}) of temporal graph

    • E.g.: -ecrate traffic-small.txt 0 10
    • Parameter description is the same to run

Data Format

  • temporal graph
    • each line consists of 4 integers "u,v,t,w"
    • u & v: starting and ending node of the edge
    • t: time stamp
    • w: weight of edge
  • interval.txt
    • each line consists of 2 integers "tb,te"
    • tb & te: beginning and ending timestamps of temporal subgraphs

Baseline & Data

  • baseline and synthetic data generator see meden_1_0_1.zip
  • KPCD groups and ground-truth see KPCD
  • BJData: available in request
  • SynData: generate by synthetic data generator (sizes up to 50G)

For any question, feel free to contact Renjun (hurenjun [AT] buaa.edu.cn)

About

"Fast Computation of Dense Temporal Subgraphs."

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages