Predictive mean matching imputation of semicontinuous variables

171 citations recorded by the source corpus.

How researchers see this paper

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

How to handle missing data: A comparison of different approaches (2015, unknown): “…research has shown that predictive mean matching is able to handle many types of data and does not directly depend on any distributional assumptions (Vink, Frank, Pannekoek, & Van Buuren, 2014) and can be used to prevent bias due to attrition in longitudinal designs (Asendorpf et al., 2014).”
Generating missing values for simulation purposes: a multivariate amputation procedure (2018, methodology): “We choose for PMM because this imputation technique samples from the observed data and thus, may be more sensitive to amputing part of the true data distribution than other imputation methods [25].”
MIDA: Multiple Imputation Using Denoising Autoencoders (2017, methodology): “et using the trained model 5: end for 3.2 Competitors and comparison Predictive mean matching (PMM) [?] and random forest (RF) [4] are the current state-of-the-art models used for multiple imputation [26, 23], PMM and RF are capable of handling mixed data types and departures from distributional ass”