Journal ArticleIdentifying Catastrophic Outlier Photometric Redshift Estimates in the COSMOS Field with Machine Learning MethodsDennis, Mitchell T (ORCID:0000000190660552); Hu, Esther M (ORCID:0009000874274617); Cowie, Lennox L (ORCID:0000000263191575)<title>Abstract</title> <p>We present the result of two binary classifier ensembled neural networks to identify catastrophic outliers for photo-<italic>z</italic>estimates within the COSMOS field utilizing only eight and five photometric bandpasses, respectively. Our neural networks can correctly classify 55.6% and 33.3% of the true positives with few to no false positives. These methods can be used to reduce the errors caused by the errors in redshift estimates, particularly at high redshift. When applied to a larger data set with only photometric data available, our eight bandpass network increased the number of objects with a photo-<italic>z</italic>greater than five from 0.1% to 1.6%, and our five bandpass network increased the number of objects with a photo-<italic>z</italic>greater than five from 0.2% to 1.8%.</p>IOP Science2025-04-1710655847The Astrophysical Journal98321730004-637Xhttps://doi.org/10.3847/1538-4357/adbe621716093High-redshift galaxiesGalaxiesNeural networksClassification1MBPDF/A1National Science Foundation