Missing and Modified Data in Nonparametric Estimation

The book gives a unified approach to nonparametric curve estimation based on missing and modified data. Missing data includes cases of missing at random and missing not at random, while data modification includes truncation and censoring, typical in survival analysis, as well as measurement errors and amplitude modulation. A universal nonparametric series E-estimator is used whose statistical idea is based on estimation of a population mean by a corresponding sample mean. While the approach is straightforward, the asymptotic theory shows that no other estimator can outperform the E-estimator.