| Título: | Semi-Supervised Affine Alignment of Manifolds |
| Autores: | Jia, Rui; Chienshiung institute technology |
| Fecha: | 2014-04-01 |
| Publicador: | TELKOMNIKA: Indonesian journal of electrical engineering |
| Fuente: |
 |
| Tipo: |
info:eu-repo/semantics/article info:eu-repo/semantics/publishedVersion |
| Tema: | manifold alignment; out of sample; affine transformation; spectral regression |
| Descripción: | Abstract: High dimensionaldata is usually produced by the source that only enjoys a limited number ofdegrees of freedom. Manifold leaning technique plays an important part infinding the correlation among the high dimensional data datasets. By making useof manifold alignment, the paired mapping relationship can be explored easily.However, the common manifold alignment algorithm can only give the mappingresult of the training set, and cannot deal with a new coming point. A newmanifold alignment algorithm is proposed in this paper. The benefit of ouralgorithm is two fold: First, the method is a semi-supervised approach, whichmakes better use of the local geometry information of the unpaired points andimproves the learning effect when the labeled proportion is very low. Second,an extended spectral aggression trick is used in the algorithm, which canproduce a linear mapping between the raw data space and the aligned space. Theexperiments result shows that, the correlation mapping can be preciselyobtained, the hidden space can be aligned effectively, and the cost of mappinga coming point is very low. |
| Idioma: | No aplica |