Bayesian Identification of Atypical Photometric Signatures

Title and Abstract: Reflect the new goal of identifying \"atypical photometric signatures\" using a \"Bayesian Non-Parametric Framework.\" Introduction: Motivates the problem of finding unusual objects within classes, rather than just classifying them. Introduces Bayesian non-parametric models as the tool. Dataset and Feature Engineering: Focuses on creating color indices (u−g,g−r,r−i,i−z) as the primary features for the Bayesian models. Methodology: This section is entirely new. It details the Bayesian Gaussian Mixture Model (BGMM) with a Dirichlet Process prior.