Two novel approaches for photometric redshift estimation based on SDSS and 2MASS databases

TL;DR

This paper compares support vector machines and Kernel Regression for estimating galaxy redshifts from SDSS and 2MASS data, highlighting their performance and optimal input parameters.

Contribution

It introduces and evaluates two novel training-set methods for photometric redshift estimation, demonstrating their superior accuracy over existing approaches.

Findings

Kernel Regression achieves rms error less than 0.02.

Support Vector Machines achieve rms error less than 0.03.

Optimal input parameters vary between methods.

Abstract

We investigate two training-set methods: support vector machines (SVMs) and Kernel Regression (KR) for photometric redshift estimation with the data from the Sloan Digital Sky Survey Data Release 5 and Two Micron All Sky Survey databases. We probe the performances of SVMs and KR for different input patterns. Our experiments show that the more parameters considered, the accuracy doesn't always increase, and only when appropriate parameters chosen, the accuracy can improve. Moreover for different approaches, the best input pattern is different. With different parameters as input, the optimal bandwidth is dissimilar for KR. The rms errors of photometric redshifts based on SVM and KR methods are less than 0.03 and 0.02, respectively. Finally the strengths and weaknesses of the two approaches are summarized. Compared to other methods of estimating photometric redshifts, they show their superiorities, especially KR, in terms of accuracy.