Automated stellar classification tools and their inter-comparison

Vasudha Bhatnagar1, Trasha Gupta2 and Ranjan Gupta3
1Department of Computer Science, University of Delhi, Delhi, India
2Inderprastha Instiute of Information Technology, Delhi, India
3IUCAA, Pune, India

View Full Article: [PDF]

Abstract

In this poster we propose a fragmented method for star classification task, which uses two different machine learning methods for predicting spectral type and luminosity class. Experiments with test data perturbed with up to 15% noise does not show significant decline in the quality of results. Comparison of the ANN predictions and predictions by the proposed method show significant correlation and share some outliers.

Fig. 1 Fig. 2 Fig. 3
   
Fig. 4    


<< Previous | Next Article >>Back to Asics_Vol_006

Keywords : stellar classification – machine learning – SMO – Random Forest – Weka