Bootstrap Bandwidth Selection for Estimating a Cosmological Mass Bias Parameter

• Activity Number: 573
• Type: Topic Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Ji Meng Loh*+
• Companies: AT&T Labs - Research
• Keywords: Bootstrap bandwidth selection ; cosmology ; two-point correlation function

Abstract

We introduce a bootstrap bandwidth selection procedure for estimating the two-point correlation function from a spatial point pattern by minimizing the integrated squared error as broken down into its variance and bias components. The variance component is estimated using nonparametric spatial bootstrap, while the bias component is estimated using a plug-in approach with a pilot estimator. We apply the method to a galaxy cluster catalog where the pilot estimator is a functional power-law model for the two-point correlation function. The resulting nonparametric estimator using the optimal bandwidth is then used to estimate a cosmological mass bias parameter.