On Nonnegative Matrix Factorization Algorithms for Signal-Dependent Noise with Application to Electromyography Data
Nonnegative matrix factorization (NMF) by the multiplicative updates algorithm is a powerful machine learning method for decomposing a high-dimensional nonnegative matrix V into two nonnegative matrices, W and H, where V ~ WH. It has been successfully applied in the analysis and interpretation of la...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
MIT Press
2015
|
| Online Access: | http://hdl.handle.net/1721.1/96302 |
| _version_ | 1811090401104232448 |
|---|---|
| author | Devarajan, Karthik Cheung, Vincent Chi-Kwan |
| author2 | Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences |
| author_facet | Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences Devarajan, Karthik Cheung, Vincent Chi-Kwan |
| author_sort | Devarajan, Karthik |
| collection | MIT |
| description | Nonnegative matrix factorization (NMF) by the multiplicative updates algorithm is a powerful machine learning method for decomposing a high-dimensional nonnegative matrix V into two nonnegative matrices, W and H, where V ~ WH. It has been successfully applied in the analysis and interpretation of large-scale data arising in neuroscience, computational biology, and natural language processing, among other areas. A distinctive feature of NMF is its nonnegativity constraints that allow only additive linear combinations of the data, thus enabling it to learn parts that have distinct physical representations in reality. In this letter, we describe an information-theoretic approach to NMF for signal-dependent noise based on the generalized inverse gaussian model. Specifically, we propose three novel algorithms in this setting, each based on multiplicative updates, and prove monotonicity of updates using the EM algorithm. In addition, we develop algorithm-specific measures to evaluate their goodness of fit on data. Our methods are demonstrated using experimental data from electromyography studies, as well as simulated data in the extraction of muscle synergies, and compared with existing algorithms for signal-dependent noise. |
| first_indexed | 2024-09-23T14:45:02Z |
| format | Article |
| id | mit-1721.1/96302 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T14:45:02Z |
| publishDate | 2015 |
| publisher | MIT Press |
| record_format | dspace |
| spelling | mit-1721.1/963022022-09-29T10:19:33Z On Nonnegative Matrix Factorization Algorithms for Signal-Dependent Noise with Application to Electromyography Data Devarajan, Karthik Cheung, Vincent Chi-Kwan Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences McGovern Institute for Brain Research at MIT Cheung, Vincent Chi-Kwan Nonnegative matrix factorization (NMF) by the multiplicative updates algorithm is a powerful machine learning method for decomposing a high-dimensional nonnegative matrix V into two nonnegative matrices, W and H, where V ~ WH. It has been successfully applied in the analysis and interpretation of large-scale data arising in neuroscience, computational biology, and natural language processing, among other areas. A distinctive feature of NMF is its nonnegativity constraints that allow only additive linear combinations of the data, thus enabling it to learn parts that have distinct physical representations in reality. In this letter, we describe an information-theoretic approach to NMF for signal-dependent noise based on the generalized inverse gaussian model. Specifically, we propose three novel algorithms in this setting, each based on multiplicative updates, and prove monotonicity of updates using the EM algorithm. In addition, we develop algorithm-specific measures to evaluate their goodness of fit on data. Our methods are demonstrated using experimental data from electromyography studies, as well as simulated data in the extraction of muscle synergies, and compared with existing algorithms for signal-dependent noise. National Institutes of Health (U.S.) (Grant NS44393) National Institutes of Health (U.S.) (Grant RC1-NS068103-01) 2015-04-01T15:27:28Z 2015-04-01T15:27:28Z 2014-05 2013-06 Article http://purl.org/eprint/type/JournalArticle 0899-7667 1530-888X http://hdl.handle.net/1721.1/96302 Devarajan, Karthik, and Vincent C. K. Cheung. “On Nonnegative Matrix Factorization Algorithms for Signal-Dependent Noise with Application to Electromyography Data.” Neural Computation 26, no. 6 (June 2014): 1128–1168. © 2014 Massachusetts Institute of Technology en_US http://dx.doi.org/10.1162/NECO_a_00576 Neural Computation Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. application/pdf MIT Press MIT Press |
| spellingShingle | Devarajan, Karthik Cheung, Vincent Chi-Kwan On Nonnegative Matrix Factorization Algorithms for Signal-Dependent Noise with Application to Electromyography Data |
| title | On Nonnegative Matrix Factorization Algorithms for Signal-Dependent Noise with Application to Electromyography Data |
| title_full | On Nonnegative Matrix Factorization Algorithms for Signal-Dependent Noise with Application to Electromyography Data |
| title_fullStr | On Nonnegative Matrix Factorization Algorithms for Signal-Dependent Noise with Application to Electromyography Data |
| title_full_unstemmed | On Nonnegative Matrix Factorization Algorithms for Signal-Dependent Noise with Application to Electromyography Data |
| title_short | On Nonnegative Matrix Factorization Algorithms for Signal-Dependent Noise with Application to Electromyography Data |
| title_sort | on nonnegative matrix factorization algorithms for signal dependent noise with application to electromyography data |
| url | http://hdl.handle.net/1721.1/96302 |
| work_keys_str_mv | AT devarajankarthik onnonnegativematrixfactorizationalgorithmsforsignaldependentnoisewithapplicationtoelectromyographydata AT cheungvincentchikwan onnonnegativematrixfactorizationalgorithmsforsignaldependentnoisewithapplicationtoelectromyographydata |