TIMERS: Error-Bounded SVD Restart on Dynamic Networks

Authors: Ziwei Zhang, Peng Cui, Jian Pei, Xiao Wang, Wenwu Zhu

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct extensive experiments on several synthetic and real dynamic networks. The experimental results demonstrate that our proposed method significantly outperforms the existing methods by reducing 27% to 42% in terms of the maximum error for dynamic network reconstruction when fixing the number of restarts.
Researcher Affiliation Academia Ziwei Zhang,1 Peng Cui,1 Jian Pei,2 Xiao Wang,1 Wenwu Zhu1 1 Department of Computer Science and Technology, Tsinghua University, China 2 School of Computing Science, Simon Fraser University, Canada
Pseudocode Yes Algorithm 1 TIMERS: Theoretically Instructed Maximum Error-bounded Restart of SVD
Open Source Code Yes 1The code is available at http://nrl.thumedia.org/
Open Datasets Yes All networks are publicly available at http://snap.stanford.edu/ or http://konect.uni-koblenz.de/
Dataset Splits No The paper describes dividing dynamic changes into 'time slices with an equal number of changes' and 'randomly hid[ing] 10% of the network' for link prediction evaluation, but it does not specify a traditional training/validation/test split for model development or hyperparameter tuning.
Hardware Specification Yes All experiments are conducted in a single PC with 2 i7-6700 CPU and 24GB memory in MATLAB language.
Software Dependencies No The paper states that experiments were conducted in "MATLAB language" but does not provide a specific version number for MATLAB or any other software dependencies with their versions.
Experiment Setup Yes Specifically, in our experiment, we set the similarity matrix to be the adjacency matrix for simplicity, and k to be 100 as commonly used. ... The other important parameter is the error threshold Θ. Qualitatively, larger Θ will tolerate more error and leads to fewer number of restarts.