Understanding Graph Sampling Algorithms for Social Network Analysis

Wang, Tianyi; Yang, Chen; Zhang, Zengbin; Xu, Tianyin; Jin, Long; Hui, Pan; Deng, Bangming; Li, Xing

doi:10.1109/icdcsw.2011.34

Cited by 78 publications

(52 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The previous studies have made it obvious that Breadth First Search (BFS) is very popular and the widely used sampling algorithm for measuring the large social networks, for example WWW or OSNs (Facebook, Twitter, etc) and helps in studying the topological characteristics such as shortest paths, clustering coefficients, node degree distribution of the sampled graphs [9]. As given in the research paper [2], the studies show that BFS has obtained a higher average clustering coefficient and very large normalised mean square error (NMSE) when compared to the other sampling algorithms due to biasing.…”

Section: Literature Reviewmentioning

confidence: 99%

“…In [4] the studies showed that this bias can be evaluated by the analysis of Markov Chain and can be improved by ReWeighted Random Walk (RWRW) and Metropolis Hastings Random Walk (MHRW) that are unbiased and this study can be justified through properties like assortivity, convergence analysis and estimation. Wang et al [2] introduced a new sampling method, Frontier Sampling (FS) based on Random Walk (RW) and claimed that FS obtains a very good degree distribution and clustering coefficient distribution when compared to MHRW and BFS.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Social network analysis primarily performs the execution in two phases: data collection and data analysis [2]. The first step includes defining the parameters and the scope of analysis.…”

Section: Introductionmentioning

confidence: 99%

“…Elicitation deals with the questionnaire/survey about specific data while registration deals with the extraction or mining of necessary information from registered entities/information. For example, MEDLINE is one of the large databases that maintain data records of the published papers on the biomedical research and had about two million records according to the statistics from 1995 to 1999 [2].…”

Section: Introductionmentioning

confidence: 99%

“…An alternative solution to this problem is Graph Sampling that provides an efficient and expensive way by selecting a representative subset of the original graph. This makes the graph scale small while keeping the characteristics of the original social graph and results in improved computations [2].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Performance Analysis and Comparison of Sampling Algorithms in Online Social Network

Stuti¹,

Srivastava²

2016

IJCA

View full text Add to dashboard Cite

Graph sampling provides an efficient way by selecting a representative subset of the original graph thus making the graph scale small for improved computations. Random walk graph sampling has been considered as a fundamental tool to collect uniform node samples from a large graph. In this paper, a comprehensive analysis and comparison of four existing sampling algorithms-BFS, NBRW-rw, MHRW and MHDA is presented. The comparison is shown on the basis of the performance of each algorithm on different kinds of datasets. Here, the considered parameters are node-degree distribution and clustering coefficient which effect the performance of an algorithm in generating unbiased samples. The sampling methods as in this study are analysed on the real-network datasets and finally the conclusion says that MHDA performs excellently whereas BFS gives a poor performance.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Section: Literature Reviewmentioning

confidence: 99%

“…Social network analysis primarily performs the execution in two phases: data collection and data analysis [2]. The first step includes defining the parameters and the scope of analysis.…”

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Performance Analysis and Comparison of Sampling Algorithms in Online Social Network

Stuti¹,

Srivastava²

2016

IJCA

View full text Add to dashboard Cite

show abstract

Sampling from social network to maintain community structure

Tong

Niu

Xie

et al. 2014

Int J Communication

View full text Add to dashboard Cite

The research of network community structure based on a large number of complex network datasets is becoming popular in recent years. For the limit of existing computing capabilities and other conditions, such a large network data processing is becoming one of the hardest issues, so sampling algorithm research has become a new hot spot in network data analysis. Based on the needs of network structure research, in this paper, we propose an improved forest fires algorithm, which can not only decrease the scale of network data but also maintain the previous network community structure well. We define two concepts, namely 'community degree' and 'center of community' in the algorithm. Then the algorithm was applied on five datasets. In order to make it convenient for the comparison between our sampling algorithm and the other six sampling algorithms under different parameters, we use network community profile and Kolmogorov-Smirno D statistics to judge the consistency between the sample and the previous graph. Experiment results show that the improved algorithm is better than the other six sampling algorithms under most of the parameters. The efficiency and feasibility of the modified algorithm is also validated. Finally, we give the recommended values of different parameters.

show abstract

Excluding pairs of tournaments

Choromański

2018

Journal of Graph Theory

View full text Add to dashboard Cite

The Erdős–Hajnal conjecture states that for every given undirected graph H there exists a constant c(H)>0 such that every graph G that does not contain H as an induced subgraph contains a clique or a stable set of size at least false|V(G)false|cfalse(Hfalse). The conjecture is still open. Its equivalent directed version states that for every given tournament H there exists a constant c(H)>0 such that every H‐free tournament T contains a transitive subtournament of order at least false|V(T)false|cfalse(Hfalse). In this article, we prove that for several pairs of tournaments, H1 and H2, there exists a constant c(H1,H2)>0 such that every {H1,H2}‐free tournament T contains a transitive subtournament of size at least false|V(T)false|cfalse(H1,H2false). In particular, we prove that for several tournaments H, there exists a constant c(H)>0 such that every {H,Hc}‐free tournament T, where Hc stands for the complement of H, has a transitive subtournament of size at least false|V(T)false|cfalse(Hfalse). To the best of our knowledge these are first nontrivial results of this type.

show abstract

Understanding Graph Sampling Algorithms for Social Network Analysis

Cited by 78 publications

References 12 publications

Performance Analysis and Comparison of Sampling Algorithms in Online Social Network

Performance Analysis and Comparison of Sampling Algorithms in Online Social Network

Sampling from social network to maintain community structure

Excluding pairs of tournaments

Contact Info

Product

Resources

About