*******************************************************************************************
Umit Cavus Buyuksahin
Strahinja Lazetic
Course: Advanced Topic of Distributed Systems
Mini Project in topic "Sampling Massive Online Graphs" that is made by Maciej Kurant
********************************************************************************************
The purpose is estimation of given offline network size. There are several techniques
that are implemented in this source code. These are:
1- Collision Counting
2- Induced Edge applied to nodes collected by
a-- Uniform Independence Sample/Without replacement (UIS WOR)
b-- Uniform Independence Sample/With Replacement (UIS WR)
c-- Weighted Independence Sample/With Replacement (WIS WR)
d-- Random Walk (RW) with Thinning/SafetyMargin
Implementations:
-GraphSize.py
it is the main code that runs all scripts and print the estimated result with
the corresponding techniques. The code includes some parameters for estimation
that are manually changed in the code:
- n_sample= number of sampled node
- teta = Thininnig factor for RW
- m = Safety Margin factor for RW
- collision.numberof_collsion(G,n_sample) = returns estimated graph size
by using collision method
- UIS_WOR.UIS_WOR(G,n_sample) = returns estimated graph size by using
UIS_WOR method
- UIS_WR.UIS_WR(A.keys(),n_sample) = returns estimated graph size by using
UIS_WR method
- Generate_Sampling_WIS.func(A,n_sample) = returns both sampled graphs for WIS
method and list of node with their degrees.
- WIS_WR.func(A,sampled_Graph,degree_Sampled_Graph) = returns estimated graph
size by using WIS_WR method
- Thinning.func(A,sampled_Graph,degree_Sampled_Graph,teta) = returns estimated
graph size by using Thinning method for random walk
- SafetyMargin.func(A,sampled_Graph,degree_Sampled_Graph,m) = returns estimated
graph size by using Safety Margin for random walk
-collision.py
it has numberof_collision(G,n_sample) function inside that takes two arguments:
- G : whole graph
- n_sample : number of sampled node
-UIS_WOR.py
it has UIS_WOR(G,n_sample) function inside that takes two arguments:
- G : whole graph
- n_sample : number of sampled node
-UIS_WR.py
it has UIS_WR(L,n_sample) function inside that takes two arguments:
- L : list of sets of neighbors
- n_sample : number of sampled node
-Generate_Sampling_WIS.py
it has func(G,n_sample) function inside that takes two arguments:
- G : whole graph
- n_sample : number of sampled node
-WIS_WR.py
it has func(G,sampled_Graph,degree_Sampled_Graph) function inside that takes four arguments:
- G : whole graph
- sampled_Graph : sampled graph
- degree_Sampled_Graph : list of degress of sampled nodes
-Thinning.py
it has func(G,sampled_Graph,degree_Sampled_Graph,teta) function inside that takes four arguments:
- G : whole graph
- sampled_Graph : sampled graph
- degree_Sampled_Graph : list of degress of sampled nodes
- teta : thinning factor
-SafetyMargin.py
it has func(A,sampled_Graph,degree_Sampled_Graph,m) function inside that takes four arguments:
- G : whole graph
- sampled_Graph : sampled graph
- degree_Sampled_Graph : list of degress of sampled nodes
- m : thinning factor
Sample Run:
When GraphSize.py script run like this:
$python GraphSize.py
all calculations will be printed.