We apply a combination of genetic algorithm (GA) and support vector machine (SVM) machine learningudalgorithms to solve two important problems faced by the astronomical community: star–galaxy separation andudphotometric redshift estimation of galaxies in survey catalogs. We use the GA to select the relevant features in theudfirst step, followed by optimization of SVM parameters in the second step to obtain an optimal set of parameters toudclassify or regress, in the process of which we avoid overfitting. We apply our method to star–galaxy separation inudPan-STARRS1 data. We show that our method correctly classifies 98% of objects down to iP1 = 24.5, with audcompleteness (or true positive rate) of 99% for galaxies and 88% for stars. By combining colors with morphology,udour star–galaxy separation method yields better results than the new SExtractor classifier spread_model, inudparticular at the faint end (iP1 > 22). We also use our method to derive photometric redshifts for galaxies in theudCOSMOS bright multiwavelength data set down to an error in (1 + z) of s = 0.013, which compares well withudestimates from spectral energy distribution fitting on the same data (s = 0.007) while making a significantly smaller number of assumptions.
展开▼
