Volume : VII, Issue : XI, November - 2018

Feature extraction based on multi-objective locally linear embedding

Han Baojin

Abstract :

 Since most of the previous data feature extraction methods use continuous feature vectors to complete feature extraction, the biggest disadvantage of this method is that after dimensionality reduction, the data does not achieve the desired dimension or even greater than the previous dimension, and thus does not actually complete the function of dimensionality reduction. The multi–objective (moea / d) [1] method is combined with the local linear embedding method, which makes the dimensionality reduction of the data take into account the actual dimensionality reduction of the data while taking into account the classification accuracy. Theoretical and experimental results show that this algorithm can effectively reduce data dimension and improve classification accuracy. < clear="all" style="page–eak–before:always;mso–eak–type:section–eak" />

Keywords :

Article: Download PDF   DOI : 10.36106/ijsr  

Cite This Article:

Feature extraction based on multi-objective locally linear embedding , Han Baojin , INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCH : Volume-7|Issue-11| November-2018

Number of Downloads : 46

References :