Smearing Estimate: A Nonparametric Retransformation Method

By N. Duan. 1983

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

A Bayesian model for repeated measures zero-inflated count data with application to outpatient psychiatric service use (2010, methodology): “, log-normal) rather than Poisson distribution for the non-zero values (Duan, 1983; Olsen and Schafer, 2001; Tooze et al., 2002).”
ROBUST ESTIMATION OF SMALL‐AREA MEANS AND QUANTILES (2010, methodology): “By combining a smearing argument (Duan, 1983) with a model for the finite-population CDF of y, CD developed a model-consistent estimator for a finite-population CDF.”
Correction factors for unbiased, efficient estimation and prediction of biomass from log-log allometric models (2013, methodology): “…allometrics, several estimates of average biomass have been commonly used, including the residual maximum likelihood (REML) estimate, also known as the Baskerville estimate (Baskerville 1972), Duan’s smearing estimate (Duan 1983), and the Snowdon correction factor or ratio estimate (Snowdon 1991).”
Predicting the popularity of online content (2008, background): “[9, 16, 21]), and in case the transformation of the dependent variable is logarithmic, the best untransformed scale estimate is”
Extending the Box–Cox transformation to the linear mixed model (2006, methodology): “Hence, retransformation methods have been proposed for univariate linear models (e.g. Duan (1983)) and may need to be studied further.”
Measuring Financial Access Around the World (2010, methodology): “There are various methods to correct for this “retransform bias”, but we choose the non-parametric smearing method (see Duan (1983)) retransformation correction which is robust to non-normal residuals (the residuals from some of our specifications do not appear to be normal based on a joint test of ”
Health insurance and the demand for medical care: evidence from a randomized experiment. (1987, unknown): “I Error term ( ) issue: Smearing estimator (Duan (1983)) Let smearing estimator φ = P exp(̂i) ni = E(exp(̂i)), where ni is the number of y > 0 and ̂i = ln(y)− xβ̂ I Expected expenditure for y > 0 E(y|y > 0, x) = E(exp(xβ̂ + ̂)) = exp(xβ̂) · E(exp(̂)) = exp(xβ̂) · φ I Expected expenditure E(y|x) =…”
Tree allometry and improved estimation of carbon stocks and balance in tropical forests (2005, methodology): “Duan (1983) suggested an even better correction procedure for back-transforming the data (see Wirth et al. 2004).”
Using Heteroscedasticity Consistent Standard Errors in the Linear Regression Model (2000, methodology): “While this approach can provide an efficient and elegant solution to the problems caused by heteroscedasticity, when the results need to be interpreted in the original scale of the variables, nonparametric methods may be necessary (Duan 1983; Carroll and Ruppert 1988:136-139).”
Bad Loans and Entry into Local Credit Markets (2004, methodology): “20 Assuming independence of the error term and the covariates, one can estimate E(p|x), where p is the default rate and x the set of regressors, using Duan’s (1983) methodology. Without this assumption an estimate of the conditional distribution of the errors must be recovered first, but this approa”