Minimally Parametric Constraints on the Primordial Power Spectrum from Lyman-alpha

TL;DR

This paper reconstructs the primordial power spectrum shape from Lyman-alpha data without assuming a simple power law, using a minimally parametric and cross-validation approach to avoid over-fitting.

Contribution

It introduces a minimally parametric method combined with cross-validation to test the primordial power spectrum shape from Lyman-alpha data, relaxing previous strong assumptions.

Findings

Reconstructed the primordial power spectrum shape from SDSS Lyman-alpha data.

Demonstrated the effectiveness of minimally parametric methods in cosmological data analysis.

Prepared for future BOSS data with improved statistical precision.

Abstract

Current analyses of the Lyman-alpha forest assume that the primordial power spectrum of density perturbations obeys a simple power law, a strong theoretical assumption which should be tested. Employing a large suite of numerical simulations which drop this assumption, we reconstruct the shape of the primordial power spectrum using Lyman-alpha data from the Sloan Digital Sky Survey (SDSS). Our method combines a minimally parametric framework with cross-validation, a technique used to avoid over-fitting the data. Future work will involve predictions for the upcoming Baryon Oscillation Sky Survey (BOSS), which will provide new Lyman-alpha data with vastly decreased statistical errors.