Unsupervised Feature Selection on Networks: A Generative View

Authors: Xiaokai Wei, Bokai Cao, Philip S. Yu

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on three real-world datasets show that our approach can select more discriminative features than state-of-the-art methods.
Researcher Affiliation Academia Xiaokai Wei , Bokai Cao and Philip S. Yu Department of Computer Science, University of Illinois at Chicago, IL, USA Institute for Data Science, Tsinghua University, Beijing, China {xwei2,caobokai,psyu}@uic.edu
Pseudocode Yes Algorithm 1 Alternating Optimization with Projected Gradient Descent
Open Source Code No The paper mentions using a K-means code available at a URL, but it does not provide access to the authors' own source code for the proposed GFS method.
Open Datasets Yes We use three publicly available network datasets with node attributes: Citeseer dataset, Cora Dataset and Wikipedia dataset 1 (Sen et al. 2008).
Dataset Splits No The paper does not explicitly provide training/test/validation dataset splits. It mentions using K-means for evaluation and repeating experiments for 20 times, but no specific splits are detailed.
Hardware Specification No The paper does not specify any hardware used for running the experiments (e.g., CPU, GPU models, memory, or cloud instances).
Software Dependencies No The paper mentions using 'K-means' code from a given URL but does not provide specific version numbers for K-means or any other software dependencies like programming languages or libraries.
Experiment Setup Yes For the proposed method GFS, we found it is not sensitive to the parameters in a reasonable range. So we fix the parameters of GFS for all datasets with β = 1 and λ = 1.