Community Detection Using Time-Dependent Personalized PageRank

Authors: Haim Avron, Lior Horesh

ICML 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experimental results suggest that it produces rankings that are distinct and competitive with the ones produced by high quality implementations of personalized Page Rank and localized heat kernel, and that our algorithm is a useful addition to the toolset of localized graph diffusions.
Researcher Affiliation Industry Haim Avron HAIMAV@US.IBM.COM IBM T. J. Watson Research Center, Yorktown Heights, NY 10598 Lior Horesh LHORESH@US.IBM.COM IBM T. J. Watson Research Center, Yorktown Heights, NY 10598
Pseudocode Yes The proposed algorithm is deterministic and simple (we give a detailed pseudo-code in the supplementary material).
Open Source Code Yes An open-source implementation of the algorithm is available through the lib Skylark library (http://xdata-skylark.github.io/libskylark/).
Open Datasets Yes Next, we evaluate the different algorithms using datasets for which ground truth communities exists (Yang & Leskovec, 2015). These are the datasets at the bottom of Table 1.
Dataset Splits No The paper describes its experimental procedure including random seed node selection and repeated runs, but it does not specify traditional training, validation, and test dataset splits in terms of percentages or sample counts for model training or tuning. Evaluation is conducted on the selected communities.
Hardware Specification No The paper does not provide any specific details regarding the hardware used for running the experiments (e.g., GPU models, CPU types, or memory specifications).
Software Dependencies No The paper mentions 'lib Skylark library' and refers to 'MATLAB' and 'CHEBFUN library'. However, it does not provide specific version numbers for any of these software components.
Experiment Setup Yes We use ϵ = 0.001. ... We run tpprgrow with ϵ = 0.001 and three different values of α: 0.85, 0.99, 1.0. We use γ = 5.0, but examine the diffusion vector at three additional time points.