SDSS STELLAR SPECTRAL CLASSIFICATION USING CONVOLUTION NEURAL NETWORKS

摘要

The present invention relates to the process of classification of objects on Sloan Digital Sky Survey (SDSS) is one of the spectroscopic surveys releasing massive datasets. Some automated techniques for analysis of such large datasets have already applied. In previous approaches, Classification has done on basis of manual feature extraction by using Feed Forward Neural Network including primitive and Non Primitive features. In Modern time various techniques have applied for such classification as Support Vector Machine, Bayesian Classifier, and Feed Forward Neural Network, KNN Method etc.