Greedy Orthogonal Pivoting Algorithm for Non-Negative Matrix Factorization
Authors: Kai Zhang, Sheng Zhang, Jun Liu, Jun Wang, Jie Zhang
ICML 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section we perform empirical evaluations. We have compared both sequential and batch GOPA with altogether 6 state-of-the-art algorithms for orthogonal NMF. |
| Researcher Affiliation | Collaboration | 1Shanghai Key Laboratory for Trustworthy Computing, School of Computer Science and Software Engineering, East China Normal University, Shanghai, China 2Infinia ML Inc., Durham, North Carolina, USA 3Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University, Shanghai, China. |
| Pseudocode | Yes | Algorithm 1 Greedy Orthogonal Pivoting Algorithm. |
| Open Source Code | Yes | Our matlab codes can be found at https://github.com/kzhang980/ORNMF. |
| Open Datasets | Yes | We have selected altogether 13 benchmark data sets that have been widely used in evaluating the performance of clustering algorithms, including sparse text data, digital images, and medical data, with details listed in Table 2. |
| Dataset Splits | No | The paper states it uses 'benchmark data sets' and 'random initialization for all the algorithms', but does not provide specific details on train/validation/test splits (e.g., percentages, sample counts, or explicit splitting methodology). |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper mentions 'Our matlab codes' but does not provide specific version numbers for Matlab or any required software libraries/dependencies. |
| Experiment Setup | Yes | We use random initialization for all the algorithms (for GOPA we use sparse random initialization namely each row of W only has one non-zero entry). To reduce statistical variations, we repeat each algorithm 30 times and report the mean and standard deviation of clustering accuracy. |