|
Activity Number:
|
154
- Some New Innovations in Survey Sampling and Missing Data Problems
|
|
Type:
|
Topic-Contributed
|
|
Date/Time:
|
Tuesday, August 10, 2021 : 10:00 AM to 11:50 AM
|
|
Sponsor:
|
Survey Research Methods Section
|
|
Abstract #317142
|
|
|
Title:
|
Random Forests Imputation in Surveys
|
|
Author(s):
|
Mehdi Dagdoug* and Camelia Goga and David Haziza
|
|
Companies:
|
Université de Bourgogne Franche Comté and Université de Bourgogne Franche Comté and University of Ottawa
|
|
Keywords:
|
survey;
missing data;
random forest;
imputation
|
|
Abstract:
|
Item nonresponse in surveys is usually handled through some form of imputation. Random forests provide flexible tools for obtaining a set of imputed values. We lay out a set of sufficient conditions needed for establishing the $L^2$-consistency of an imputed estimator based on random forests. We consider several random forests algorithms and establish the $L^2$-consistency for each algorithm. We derive the asymptotic variance of imputed estimators and propose a consistent variance estimator. We present the results from a simulation study that investigates the performance of point and variance estimators based on random forest imputation in terms of bias, efficiency and coverage rate.
|
Authors who are presenting talks have a * after their name.
