Selección automática de parámetros en LLE / Automatic selection of parameters in LLE
In: Revista Facultad de Ingeniería Universidad de Antioquia, 2010, Heft 56, S. 170-181
academicJournal
- print, 11 ref
Zugriff:
Locally Linear Embedding (LLE) is a nonlinear dimensionality reduction technique, which preserves the local geometry of high dimensional space performing an embedding to low dimensional space. LLE algorithm has 3 free parameters that must be set to calculate the embedding: the number of nearest neighbors κ, the output space dimensionality m and the regularization parameter α. The last one only is necessary when the value of κ is greater than the dimensionality of input space or data are not located in general position, and it plays an important role in the embedding results. In this paper we propose a pair of criteria to find the optimum value for the parameters κ and α, to obtain an embedding that faithfully represent the input data space. Our approaches are tested on 2 artificial data sets and 2 real world data sets to verify the effectiveness of the proposed criteria, besides the results are compared against methods found in the state of art.
Titel: |
Selección automática de parámetros en LLE / Automatic selection of parameters in LLE
|
---|---|
Autor/in / Beteiligte Person: | VALENCIA AGUIRRE, Juliana ; ALVAREZ MEZA, Andrés Marino ; DAZA SANTACOLOMA, Genaro ; ACOSTA MEDINA, Carlos Daniel ; CASTELLANOS DOMINGUEZ, Germán |
Link: | |
Zeitschrift: | Revista Facultad de Ingeniería Universidad de Antioquia, 2010, Heft 56, S. 170-181 |
Veröffentlichung: | Medellín: Facultad de Ingeniería. Universidad de Antioquia, 2010 |
Medientyp: | academicJournal |
Umfang: | print, 11 ref |
ISSN: | 0120-6230 (print) |
Schlagwort: |
|
Sonstiges: |
|