Redshift is a key quantity for inferring cosmological model parameters. In photometric redshift estimation, cosmologists use the coarse data collected from the vast majority of galaxies to predict the redshift of individual galaxies. To properly quantify the uncertainty in the predictions, however, one needs to go beyond standard regression and instead estimate the full conditional density $f(z|\mathbf{x})$ of a galaxy’s redshift $z$ given its photometric covariates $\mathbf{x}$. The problem is further complicated by selection bias: usually only the rarest and brightest galaxies have known redshifts, and these galaxies have characteristics and measured covariates that do not necessarily match those of more numerous and dimmer galaxies of unknown redshift. Unfortunately, there is not much research on how to best estimate complex multivariate densities in such settings.
Here we describe a general framework for properly constructing and assessing nonparametric conditional density estimators under selection bias, and for combining two or more estimators for optimal performance. We propose new improved photo-$z$ estimators and illustrate our methods on data from the Sloan Data Sky Survey and an application to galaxy–galaxy lensing. Although our main application is photo-$z$ estimation, our methods are relevant to any high-dimensional regression setting with complicated asymmetric and multimodal distributions in the response variable.
References
Cunha, C. E., Lima, M., Oyaizu, H., Frieman, J. and Lin, H. (2009). Estimating the redshift distribution of photometric galaxy samples—II. Applications and tests of a new method. Mon. Not. R. Astron. Soc. 396 2379–2398.Cunha, C. E., Lima, M., Oyaizu, H., Frieman, J. and Lin, H. (2009). Estimating the redshift distribution of photometric galaxy samples—II. Applications and tests of a new method. Mon. Not. R. Astron. Soc. 396 2379–2398.
Dahlen, T., Mobasher, B., Faber, S. M., Ferguson, H. C., Barro, G., Finkelstein, S. L., Finlator, K., Fontana, A., Gruetzbauch, R., Johnson, S. et al. (2013). A critical assessment of photometric redshift methods: A CANDELS investigation. Astrophys. J. 775 93.Dahlen, T., Mobasher, B., Faber, S. M., Ferguson, H. C., Barro, G., Finkelstein, S. L., Finlator, K., Fontana, A., Gruetzbauch, R., Johnson, S. et al. (2013). A critical assessment of photometric redshift methods: A CANDELS investigation. Astrophys. J. 775 93.
Gretton, A., Smola, A., Huang, J., Schmittfull, M., Borgwardt, K. and Schölkopf, B. (2010). Covariate shift by kernel mean matching. In Dataset Shift in Machine Learning (J. Quionero-Candela, M. Sugiyama, A. Schwaighofer and N. D. Lawrence, eds.) Chapter 8. MIT Press, Cambridge, MA.Gretton, A., Smola, A., Huang, J., Schmittfull, M., Borgwardt, K. and Schölkopf, B. (2010). Covariate shift by kernel mean matching. In Dataset Shift in Machine Learning (J. Quionero-Candela, M. Sugiyama, A. Schwaighofer and N. D. Lawrence, eds.) Chapter 8. MIT Press, Cambridge, MA.
Hastie, T., Tibshirani, R. and Friedman, J. H. (2009). The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer, Berlin.Hastie, T., Tibshirani, R. and Friedman, J. H. (2009). The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer, Berlin.
Izbicki, R. and Lee, A. B. (2016). Nonparametric conditional density estimation in a high-dimensional regression setting. J. Comput. Graph. Statist. 25 1297–1316.Izbicki, R. and Lee, A. B. (2016). Nonparametric conditional density estimation in a high-dimensional regression setting. J. Comput. Graph. Statist. 25 1297–1316.
Izbicki, R., Lee, A. B and Freeman, P. E (2017). Supplement to “Photo-$z$ estimation: An example of nonparametric conditional density estimation under selection bias.” DOI:10.1214/16-AOAS1013SUPP.Izbicki, R., Lee, A. B and Freeman, P. E (2017). Supplement to “Photo-$z$ estimation: An example of nonparametric conditional density estimation under selection bias.” DOI:10.1214/16-AOAS1013SUPP.
Izbicki, R., Lee, A. B. and Schafer, C. M. (2014). High-dimensional density ratio estimation with extensions to approximate likelihood computation. J. Mach. Learn. Res. 33.Izbicki, R., Lee, A. B. and Schafer, C. M. (2014). High-dimensional density ratio estimation with extensions to approximate likelihood computation. J. Mach. Learn. Res. 33.
Kanamori, T., Suzuki, T. and Sugiyama, M. (2012). Statistical analysis of kernel-based least-squares density-ratio estimation. Mach. Learn. 86 335–367.Kanamori, T., Suzuki, T. and Sugiyama, M. (2012). Statistical analysis of kernel-based least-squares density-ratio estimation. Mach. Learn. 86 335–367.
Kind, M. C. and Brunner, R. J. (2013). Tpz: Photometric redshift pdfs and ancillary information by using prediction trees and random forests. Mon. Not. R. Astron. Soc. 432 1483–1501.Kind, M. C. and Brunner, R. J. (2013). Tpz: Photometric redshift pdfs and ancillary information by using prediction trees and random forests. Mon. Not. R. Astron. Soc. 432 1483–1501.
Kremer, J., Gieseke, F., Pedersen, K. S. and Igel, C. (2015). Nearest neighbor density ratio estimation for large-scale applications in astronomy. Astron. Comput.Kremer, J., Gieseke, F., Pedersen, K. S. and Igel, C. (2015). Nearest neighbor density ratio estimation for large-scale applications in astronomy. Astron. Comput.
Lima, M., Cunha, C. E., Oyaizu, H., Frieman, J., Lin, H. and Sheldon, E. (2008). Estimating the redshift distribution of photometric galaxy samples. Mon. Not. R. Astron. Soc. 390 118–130.Lima, M., Cunha, C. E., Oyaizu, H., Frieman, J., Lin, H. and Sheldon, E. (2008). Estimating the redshift distribution of photometric galaxy samples. Mon. Not. R. Astron. Soc. 390 118–130.
Mandelbaum, R., Seljak, U., Hirata, C. M., Bardelli, S., Bolzonella, M., Bongiorno, A., Carollo, M., Contini, T., Cunha, C. E., Garilli, B., Iovino, A., Kampczyk, P., Kneib, J. P., Knobel, C., Koo, D. C., Lamareille, F., Le Fevre, O., Leborgne, J. F., Lilly, S. J., Maier, C., Mainieri, V., Mignoli, M., Newman, J. A., Oesch, P. A., Perez-Montero, E., Ricciardelli, E., Scodeggio, M., Silverman, J. and Tasca, L. (2008). Precision photometric redshift calibration for galaxy–galaxy weak lensing. Mon. Not. R. Astron. Soc. 386 781–806.Mandelbaum, R., Seljak, U., Hirata, C. M., Bardelli, S., Bolzonella, M., Bongiorno, A., Carollo, M., Contini, T., Cunha, C. E., Garilli, B., Iovino, A., Kampczyk, P., Kneib, J. P., Knobel, C., Koo, D. C., Lamareille, F., Le Fevre, O., Leborgne, J. F., Lilly, S. J., Maier, C., Mainieri, V., Mignoli, M., Newman, J. A., Oesch, P. A., Perez-Montero, E., Ricciardelli, E., Scodeggio, M., Silverman, J. and Tasca, L. (2008). Precision photometric redshift calibration for galaxy–galaxy weak lensing. Mon. Not. R. Astron. Soc. 386 781–806.
Margolis, A. (2011). A literature review of domain adaptation with unlabeled data. Available at http://ssli.ee.washington.edu/~amargoli/review_Mar23.pdf.Margolis, A. (2011). A literature review of domain adaptation with unlabeled data. Available at http://ssli.ee.washington.edu/~amargoli/review_Mar23.pdf.
Minh, H. Q., Niyogi, P. and Yao, Y. (2006). Mercer’s theorem, feature maps, and smoothing. In Learning Theory, 19th Annual Conference on Learning Theory.Minh, H. Q., Niyogi, P. and Yao, Y. (2006). Mercer’s theorem, feature maps, and smoothing. In Learning Theory, 19th Annual Conference on Learning Theory.
Moreno-Torres, J. G., Raeder, T., Alaíz-Rodríguez, R., Chawla, N. V. and Herrera, F. (2012). A unifying view on dataset shift in classification. Pattern Recogn. 45 521–530.Moreno-Torres, J. G., Raeder, T., Alaíz-Rodríguez, R., Chawla, N. V. and Herrera, F. (2012). A unifying view on dataset shift in classification. Pattern Recogn. 45 521–530.
Sheldon, E. S., Cunha, C. E., Mandelbaum, R., Brinkmann, J. and Weaver, B. A. (2012). Photometric redshift probability distributions for galaxies in the SDSS DR8. Astrophys. J., Suppl. Ser. 201 32.Sheldon, E. S., Cunha, C. E., Mandelbaum, R., Brinkmann, J. and Weaver, B. A. (2012). Photometric redshift probability distributions for galaxies in the SDSS DR8. Astrophys. J., Suppl. Ser. 201 32.
Shimodaira, H. (2000). Improving predictive inference under covariate shift by weighting the log-likelihood function. J. Statist. Plann. Inference 90 227–244.Shimodaira, H. (2000). Improving predictive inference under covariate shift by weighting the log-likelihood function. J. Statist. Plann. Inference 90 227–244.
Sugiyama, M., Suzuki, T., Nakajima, S., Kashima, H., von Bünau, P. and Kawanabe, M. (2008). Direct importance estimation for covariate shift adaptation. Ann. Inst. Statist. Math. 60 699–746.Sugiyama, M., Suzuki, T., Nakajima, S., Kashima, H., von Bünau, P. and Kawanabe, M. (2008). Direct importance estimation for covariate shift adaptation. Ann. Inst. Statist. Math. 60 699–746.
Weiner, B. J., Phillips, A. C., Faber, S. M., Willmer, C. N. A., Vogt, N. P. et al. (2005). The DEEP groth strip galaxy redshift survey. III. Redshift catalog and properties of galaxies. Astrophys. J. 620 595.Weiner, B. J., Phillips, A. C., Faber, S. M., Willmer, C. N. A., Vogt, N. P. et al. (2005). The DEEP groth strip galaxy redshift survey. III. Redshift catalog and properties of galaxies. Astrophys. J. 620 595.
York, D. G., Adelman, J., Anderson, J. E. Jr., Anderson, S. F., Annis, J., Bahcall, N. A., Bakken, J. A., Barkhouser, R., Bastian, S., Berman, E. et al. (2000). The Sloan Digital Sky Survey: Technical summary. Astrophys. J. 120 1579.York, D. G., Adelman, J., Anderson, J. E. Jr., Anderson, S. F., Annis, J., Bahcall, N. A., Bakken, J. A., Barkhouser, R., Bastian, S., Berman, E. et al. (2000). The Sloan Digital Sky Survey: Technical summary. Astrophys. J. 120 1579.
Zhao, L. C. and Liu, Z. J. (1985). Strong consistency of the kernel estimators of conditional density function. Acta Math. Appl. Sin. Engl. Ser. 1 314–318.Zhao, L. C. and Liu, Z. J. (1985). Strong consistency of the kernel estimators of conditional density function. Acta Math. Appl. Sin. Engl. Ser. 1 314–318.
