On classification of signals represented with data-dependent overcomplete dictionaries
| العنوان: | On classification of signals represented with data-dependent overcomplete dictionaries |
|---|---|
| المؤلفون: | Nicola Ancona, Rosalia Maglietta |
| المصدر: | International Journal of General Systems. 40:854-882 |
| بيانات النشر: | Informa UK Limited, 2011. |
| سنة النشر: | 2011 |
| مصطلحات موضوعية: | Generalization, business.industry, Supervised learning, Pattern recognition, Sparse approximation, External Data Representation, Computer Science Applications, Theoretical Computer Science, Support vector machine, symbols.namesake, Control and Systems Engineering, Modeling and Simulation, Kernel (statistics), Gaussian function, symbols, Artificial intelligence, Representation (mathematics), business, Information Systems, Mathematics |
| الوصف: | This paper focuses on the problem of how data representation influences the generalization error of kernel-based learning machines like support vector machines (SVMs). We analyse the effects of sparse and dense data representations on the generalization error of SVM. We show that using sparse representations the performances of classifiers belonging to hypothesis spaces induced by polynomial or Gaussian kernel functions reduce to the performances of linear classifiers. Sparse representations reduce the generalization error as long as the representation is not too sparse as with very large dictionaries. Dense data representations reduce the generalization error also using very large dictionaries. We use two schemes for representing data in data-independent overcomplete Haar and Gabor dictionaries, and measure the generalization error of SVMs on benchmark datasets. We study sparse and dense representations in the case of data-dependent overcomplete dictionaries and we show how this leads to principal compon... |
| تدمد: | 1563-5104 0308-1079 |
| الوصول الحر: | https://explore.openaire.eu/search/publication?articleId=doi_________::35c38c7da67686230438e95b19996185Test https://doi.org/10.1080/03081079.2011.622100Test |
| رقم الانضمام: | edsair.doi...........35c38c7da67686230438e95b19996185 |
| قاعدة البيانات: | OpenAIRE |
| تدمد: | 15635104 03081079 |
|---|