Predicting conditional probability distributions of redshifts of Active Galactic Nuclei using Hierarchical Correlation Reconstruction
This work addresses the practical challenge of redshift prediction in astrophysics, offering an incremental improvement by extending HCR with Canonical Correlation Analysis and l1 regularization for feature optimization.
The paper tackles the problem of predicting complex conditional probability distributions, such as multimodal ones, for Active Galactic Nuclei redshifts using the Hierarchical Correlation Reconstruction approach, achieving interpretable models with coefficients that describe feature contributions to conditional moments.
While there is a general focus on prediction of values, real data often only allows to predict conditional probability distributions, with capabilities bounded by conditional entropy $H(Y|X)$. If additionally estimating uncertainty, we can treat a predicted value as the center of Gaussian of Laplace distribution - idealization which can be far from complex conditional distributions of real data. This article applies Hierarchical Correlation Reconstruction (HCR) approach to inexpensively predict quite complex conditional probability distributions (e.g. multimodal): by independent MSE estimation of multiple moment-like parameters, which allow to reconstruct the conditional distribution. Using linear regression for this purpose, we get interpretable models: with coefficients describing contributions of features to conditional moments. This article extends on the original approach especially by using Canonical Correlation Analysis (CCA) for feature optimization and l1 "lasso" regularization, focusing on practical problem of prediction of redshift of Active Galactic Nuclei (AGN) based on Fourth Fermi-LAT Data Release 2 (4LAC) dataset.