Nonparametric reconstruction of the cosmic expansion with local regression smoothing and simulation extrapolation

Understand

In this work we present a nonparametric approach, which works on minimal assumptions, to reconstruct the cosmic expansion of the Universe.

• We propose to combine a locally weighted scatterplot smoothing method and a simulation-extrapolation method.
• The first one (Loess) is a nonparametric approach that allows to obtain smoothed curves with no prior knowledge of the functional relationship between variables nor of the cosmological quantities.
• The second one (Simex) takes into account the effect of measurement errors on a variable via a simulation process.