Aims: We present a custom support vector machine classification package for photometric redshift estimation, including comparisons with other methods. We also explore the efficacy of the inclusion of galaxy shape information in redshift estimation. Support vector machine algorithms, a type of machine learning, utilize statistical learning theory and optimization theory to construct predictive models based on the information content of data in a way that can treat different input types symmetrically, which can be a useful estimator of the additional information contained in parameters, such as those describing the morphology of the galaxies. Methods: The code we have developed is designated SPIDERz and is made available to the community. As test data for evaluating performance and comparison with other methods we use 1) the publically available portion of the PHAT-1 catalog based on the GOODS-N field with redshifts ranging from $0.08 < z < 3.6$, 2) 14365 galaxies from the COSMOS bright survey with photometry, photometric magnitudes, and spectroscopic redshifts, which are in the redshift range $z < 1.4$, and 3) imaging and five band photometric magnitudes from the All-wavelength Extended Groth Strip International Survey with redshifts $0.01 < z < 1.57$. Results: We find that SPIDER-z achieves results competitive with other empirical packages on the PHAT-1 data, and performs relatively well in estimating redshifts with the COSMOS and AEGIS data. We determine from analyses with both the COSMOS and AEGIS data that the inclusion of morphological information does not have a statistically significant benefit for photometric redshift estimation with the techniques employed here. We conclude that it is likely a generic result for empirical redshift estimation techniques that the inclusion of morphological information does not improve metrics such as the RMS error or number of outliers.
↧