Unbiased Multivariate Correlation Analysis

Authors: Yisen Wang, Simone Romano, Vinh Nguyen, James Bailey, Xingjun Ma, Shu-Tao Xia

AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments In this section, we empirically assess UMC. Firstly, we study UMC s performance on synthetic data. Secondly, we incorporate UMC into a state-of-the-art subspace beam search algorithm (Keller, Muller, and Bohm 2012) to mine correlated subspaces for subspace clustering and outlier detection on real data.
Researcher Affiliation Academia Dept. of Computer Science and Technology, Tsinghua University, China Dept. of Computing and Information Systems, University of Melbourne, Australia
Pseudocode No The paper does not include any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper mentions a link to 'Supplementary Material' for proofs ('https://sites.google.com/site/umcsupplementary/home') but does not explicitly state that the source code for the methodology is available at this or any other location.
Open Datasets Yes We test 11 real UCI data sets2 widely used for benchmarking in the clustering community and previous work on correlation measures, using their class labels as ground truth. 2http://archive.ics.uci.edu/ml/index.html
Dataset Splits No The paper mentions using synthetic data of a certain size and real-world datasets, but it does not provide specific training, validation, or test split percentages or counts needed for reproduction.
Hardware Specification Yes We examine the scalability of the measures with regards to dimensionality and data size on a PC platform with Intel Core i7-3770 CPU and 32GB RAM.
Software Dependencies No The paper describes the computational aspects and methods used but does not provide specific software dependencies with version numbers (e.g., library names with versions).
Experiment Setup Yes For UMC, we set ϵ = 0.3.