A Convex Formulation for Spectral Shrunk Clustering
Authors: Xiaojun Chang, Feiping Nie, Zhigang Ma, Yi Yang, Xiaofang Zhou
AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | To validate the efficacy of the proposed algorithm, we perform extensive experiments on several benchmark datasets in comparison with some state-of-the-art clustering approaches. The experimental results demonstrate that the proposed algorithm has quite promising clustering performance. Experiment In this section, we perform extensive experiments on a variety of applications to test the performance of our method SSC. |
| Researcher Affiliation | Academia | 1Centre for Quantum Computation and Intelligent Systems, University of Technology Sydney, Australia. 2Center for OPTical IMagery Analysis and Learning, Northwestern Polytechnical University, Shaanxi, China. 3Department of Computer Science and Engineering, University of Texas at Arlington, USA. 4School of Computer Science, Carnegie Mellon University, USA. 5School of Information Technology & Electrical Engineering, The University of Queensland, Australia. |
| Pseudocode | Yes | Algorithm 1: Optimization Algorithm for SSC |
| Open Source Code | No | The paper does not provide any explicit statements about open-sourcing the code or links to a code repository. |
| Open Datasets | Yes | A variety of datasets are used in our experiments which are described as follows. The AR dataset (Martinez and Benavente 1998) contains 840 faces of 120 different people. The JAFFE dataset (Lyons et al. 1997) consists of 213 images of different facial expressions from 10 different Japanese female models. The ORL dataset (Samaria and Harter 1994) consists of 40 different subjects with 10 images each. The UMIST face dataset (Graham and M 1998) consists of 564 images of 20 individuals with mixed race, gender and appearance. The Bin Alpha dataset contains 26 binary hand-written alphabets and we randomly select 30 images for every alphabet. The MSRA50 dataset contains 1799 images from 12 different classes. The Yale B dataset (Georghiades, Belhumeur, and Kriegman 2001) contains 2414 near frontal images from 38 persons under different illuminations. We additionally use the USPS dataset to validate the performance on handwritten digit recognition. |
| Dataset Splits | No | The paper mentions repeating clustering 50 times with random initialization and reporting best results, and projecting data into a low-dimensional subspace, but does not provide specific train/validation/test dataset splits or cross-validation details. |
| Hardware Specification | No | The paper does not provide any specific hardware details such as CPU/GPU models, memory specifications, or types of computing resources used for the experiments. |
| Software Dependencies | No | The paper does not specify any software dependencies with version numbers, such as programming languages, libraries, or frameworks used for implementation. |
| Experiment Setup | Yes | The size of neighborhood, k is set to 5 for all the spectral clustering algorithms. For parameters in all the comparison algorithms, we tune them in the range of {10^-6, 10^-3, 10^0, 10^3, 10^6} and report the best results. The parameter γ is fixed at 1, which is the median value of the tuned range of the parameters. |