Speaker
Description
The internal composition of compact stars remains largely unresolved in spite of the latest mass–radius measurements by NICER and tidal deformability constraints from gravitational-wave observations. A central challenge is to identify the distinction in the macroscopic observable structure of nucleonic neutron stars and self-bound quark stars. We employ a leave-one-out cross-validation approach in a Bayesian analysis. We quantify the agreement of individual stellar observations with nucleonic and quark equation of state (EoS) models satisfying established theoretical and experimental constraints using posterior pointwise predictive density. Our analysis shows that heavier pulsars, such as PSR J0740+6620, and low-mass objects like HESS J1731–347, are better described by quark-matter EoSs, while stars near the canonical 1.4 M$_\odot$ mass exhibit ambiguous behavior. PSR J0030+0451 shows marginal preference for nucleonic matter, whereas others lean toward quark models. These results highlight the sensitivity of model selection to individual observations and the need for more precise measurements. The methodology is readily extendable to hybrid and hyperonic star models, among other exotic possibilities, offering a systematic framework for probing the preferences of the observational data. Unlike global EoS inference, our approach isolates how each observation individually prefers or challenges competing models.