Subsampling (Springer Series in Statistics)

Rating:
Summary: Theoretical treatise on subsampling
Review: Subsampling is a form of data reuse similar to the bootstrap. However the idea of subsampling preceded the bootstrap. The original work is due to Hartigan in the 1970s and was motivation for some of Efron's ideas on bootstrap. The authors begin with a treatment of bootstrap theory. They then introduce subsampling under the assumption that the data are independent and identically distributed (i.i.d). They then provide the theoretical development of subsampling for i.i.d. data, stationary time series, non-stationary time series, random fields and marked point processes. The material is thorough and includes some of the authors' own original research. It is a highly advanced text that presupposes a first year graduate course in theoretical statistics.