Crawling the Community Structure of Multiplex Networks

Authors: Ricky Laishram, Jeremy D. Wendt, Sucheta Soundarajan168-175

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We test MCS against six baseline algorithms on real-world multiplex networks, and achieved large gains in performance.
Researcher Affiliation Collaboration Ricky Laishram Syracuse University Syracuse NY, USA rlaishra@syr.edu Jeremy D. Wendt Sandia National Laboratories Albuquerque, NM, USA jdwendt@sandia.gov Sucheta Soundarajan Syracuse University Syracuse NY, USA susounda@syr.edu
Pseudocode No The paper describes the algorithm in text, but does not provide a formal pseudocode or algorithm block.
Open Source Code No The paper does not provide a specific repository link or explicit statement about the release of their own source code for the MCS methodology.
Open Datasets Yes The networks we use for our experiments are given in Table 2. The Twitter networks are collected using the Twitter API (Wendt et al. 2016). These networks contain 3 layers, and approximately 2k nodes and 10k edges. For the co-authorship network6, ca Hep Th is considered as the expensive layer, and ca Hep Ph is considered as the cheap layer. This network contains approximately 1.3k nodes and 1.9k edges. The Noordin Top network7 has five layers... The last network we use is the DBLP dataset8 between 2013 and 2017. ...6https://snap.stanford.edu/data/ 7https://sites.google.com/site/sfeverton18/research/appendix-1 8https://dblp.uni-trier.de/
Dataset Splits No The paper does not explicitly provide training/test/validation dataset splits or mention a specific validation set.
Hardware Specification No The paper only states that computational resources were provided by Syracuse University, without specifying any hardware details like GPU/CPU models, processors, or memory.
Software Dependencies No The paper mentions the use of the Louvain community detection method, but does not specify any software dependencies with version numbers.
Experiment Setup Yes Experimental Setup We use the Louvain community detection method (Blondel et al. 2011)9. The budget and the layer costs are given in Table 2. We perform 10 trials of each experiment to account for randomness in sampling as well as in the Louvain method. We also keep the ordering of the nodes consistent to get rid of the effect of node orders during the community detection process. For UQR, (βµ x, βσ x) of all layers in the Twitter, coauthorship and DBLP networks are set to (0.2, 0.1). For the Noordin Top network, it is set to (0.2, 0.1) (communication), (0.5, 0.1) (friendship), (0.25, 0.1) (mentors), and (0.75, 0.1) (kinship and classmates).