<abstract xmlns="http://www.w3.org/1999/xhtml"><p> ReGenesees is a new software system for design-based and model-assisted analysis of complex sample surveys, based on R. As compared to traditional estimation platforms, it ensures easier and safer usage and achieves a dramatic reduction in user workload for both the calibration and the variance estimation tasks. Indeed, ReGenesees allows the specification of calibration models in a symbolic way, using R model formulae. Driven by this symbolic metadata, the system automatically and transparently generates the right values and formats for the auxiliary variables at the sample level, and assists the user in defining and calculating the corresponding population totals. Moreover, ReGenesees can handle arbitrary complex estimators, provided they can be expressed as differentiable functions of Horvitz-Thompson or calibration estimators of totals. Complex estimators can be defined in a completely free fashion: the user only needs to provide the system with the symbolic expression of the estimator as a mathematical function. ReGenesees is in fact able to automatically linearize such complex estimators, so that the estimation of their variance comes at no cost at all to the user. Remarkably, all the innovative features sketched above leverage a particular strong point of the R programming language, namely its ability to process symbolic information.</p></abstract>

ReGenesees is a new software system for design-based and model-assisted analysis of complex sample surveys, based on R. As compared to traditional estimation platforms, it ensures easier and safer usage and achieves a dramatic reduction in user workload for both the calibration and the variance estimation tasks. Indeed, ReGenesees allows the specification of calibration models in a symbolic way, using R model formulae. Driven by this symbolic metadata, the system automatically and transparently generates the right values and formats for the auxiliary variables at the sample level, and assists the user in defining and calculating the corresponding population totals. Moreover, ReGenesees can handle arbitrary complex estimators, provided they can be expressed as differentiable functions of Horvitz-Thompson or calibration estimators of totals. Complex estimators can be defined in a completely free fashion: the user only needs to provide the system with the symbolic expression of the estimator as a mathematical function. ReGenesees is in fact able to automatically linearize such complex estimators, so that the estimation of their variance comes at no cost at all to the user. Remarkably, all the innovative features sketched above leverage a particular strong point of the R programming language, namely its ability to process symbolic information.

ReGenesees: an Advanced R System for Calibration, Estimation and Sampling Error Assessment in Complex Sample Surveys

Istat – Italian National Institute of Statistics, Via Cesare Balbo, 16 Rome, Rome 00184, Italy

Journal of Official Statistics

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

ReGenesees is a new software system for design-based and model-assisted analysis of complex sample surveys, based on R. As compared to traditional estimation...

ReGenesees: an Advanced R System for Calibration, Estimation and Sampling Error Assessment in Complex Sample Surveys

Publié en ligne: 27 juin 2015

Pages: 177 - 203

Reçu: 01 juil. 2013

Accepté: 01 août 2014

DOI: https://doi.org/10.1515/jos-2015-0013

Mots clés
Complex estimators, variance estimation, automated linearization, symbolic computation

© by Diego Zardetto

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

ReGenesees: an Advanced R System for Calibration, Estimation and Sampling Error Assessment in Complex Sample Surveys

Publié en ligne: 27 juin 2015

Pages: 177 - 203

Reçu: 01 juil. 2013

Accepté: 01 août 2014

DOI: https://doi.org/10.1515/jos-2015-0013

Mots clésComplex estimators, variance estimation, automated linearization, symbolic computation

© by Diego Zardetto

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Mots clés
Complex estimators, variance estimation, automated linearization, symbolic computation