Self-paced Consensus Clustering with Bipartite Graph
Authors: Peng Zhou, Liang Du, Xuejun Li
IJCAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The extensive experimental results demonstrate the effectiveness and superiority of the proposed method. |
| Researcher Affiliation | Academia | 1School of Computer Science and Technology, Anhui University 2State Key Laboratory of Computer Science, Institute of Software, Chinese Academy of Sciences 3School of Computer and Information Technology, Shanxi University |
| Pseudocode | Yes | Algorithm 1 SCCBG |
| Open Source Code | No | The paper does not provide an explicit statement or link for open-source code for the described methodology. |
| Open Datasets | Yes | We use 8 data sets, including ALLAML1, GLIOMA1, K1b [Zhao and Karypis, 2004], Lung1, Medical [Zhou et al., 2015b], Tdt22, Tr41 [Zhao and Karypis, 2004], and TOX1. |
| Dataset Splits | No | The paper describes how base clusterings are generated and how results are averaged over subsets of these base clusterings, but it does not specify a training/validation/test split for the datasets used in the experiments. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., CPU/GPU models, memory) used for running experiments. |
| Software Dependencies | No | The paper mentions using 'quadprog function provided in Matlab' but does not specify the version of Matlab or any other software dependencies with version numbers. |
| Experiment Setup | Yes | Following the experimental setup in [Wang et al., 2009b; Zhou et al., 2015b], we use k-means to generate the base clusterings. In more detail, we run k-means 200 times with different initializations to obtain 200 base results. Then we divide them into 10 subsets, with 20 base results in each subset. Next, we apply consensus clustering methods on each subsets, and report the average results on the 10 subsets. ... Our method adjusts λ automatically as introduced in Algorithm 1. The parameter ρ is also automatically decided. We first initialize ρ = 1, and then, if the rank of L is larger than n + k c, we double it. If its rank is smaller than n + k c, we reduce ρ by half. The only hyper-parameter needed to tune manually is γ. As discussed before, γ should not be too large to make the subproblem convex, thus we tune it in the range [10 5, 100]. |