R packages
wbacon
Weighted BACON algorithms for multivariate outlier nomination (detection) and robust linear regression
Learn moreUnpackaged software (languages: C, R, Fortran, and JS)
Robust self-calibration
- Topic
- Microsimulation
- Description
- The method of robust self-calibration attempts to calibrate the data (or, equivalently, the sampling weights) such that the Horvitz-Thompson estimator of the total is aligned with the known population total (likewise, alignment of the estimated weighted sample mean with the known population mean). The method is robust in the sense that the influence of outliers or influential observations is controlled.
- Reference
- Schoch, T. and A. Müller, 2020. Treatment of sample under-representation and skewed heavy-tailed distributions in survey-based microsimulation: An analysis of redistribution effects in compulsory health care insurance in Switzerland, AStA Wirtschafts- und Sozialstatistisches Archiv 14 (3), pp. 267-304, Link
- License
- GPL >= 2 (Tobias Schoch, vers. 0.1, July 9, 2020)
- Dependencies
- R package robsurvey
- Download
- selfcalibrate_rob.R
Robust minimum estimated risk M-estimation/calibration
- Topic
- Microsimulation
- Description
- The method of robust minimum estimated risk M-estimation/calibration (MR estimator) seeks alignment of a robustified Horvitz-Thompson estimator with the known population total (or the population mean); in this respect, the method is comparable with robust self-calibration. However, the MR estimator can achieve higher efficiency than robust self-calibration.
- Reference
- Schoch, T. and A. Müller, 2020. Treatment of sample under-representation and skewed heavy-tailed distributions in survey-based microsimulation: An analysis of redistribution effects in compulsory health care insurance in Switzerland, AStA Wirtschafts- und Sozialstatistisches Archiv 14 (3), pp. 267-304, Link
- License
- GPL >= 2 (Tobias Schoch, vers. 0.1, July 9, 2020)
- Dependencies
- R package robsurvey
- Download
- mr.R
Weighted quantile
- Topic
- Weighted quantile (survey sampling)
- Description
- Weighted quantile and k-th order statistic
- Language
- C99
- License
- GPL >= 2 (Tobias Schoch, September 14, 2017)
- Download
- Link to GitHub
kNN imputation
- Topic
- Imputation (survey sampling)
- Description
- Hotdeck imputation by the method of k nearest neighbors (random imputation and average)
- Reference
- Schoch, T. and B. Hulliger, 2019. Evaluation der Einsetzungsqualität im Vermögensmodul von SILC, Studie im Auftrag des Bundesamts für Statistik, Olten
- License
- GPL >= 2 (Tobias Schoch, vers. 0.1, September 26, 2018)
- Dependencies
- R package gower
- Download
- kNN.R
Weighted iterative proportional fitting (IPF)
- Topic
- (Micro-) simulation
- Description
- Implements the IPF method of Deming and Stephan (1940) and a weighted approach (see documentation)
- Reference
- Müller, A., C. Lieb, and T. Schoch, 2016. Räumliche Entwicklung der Arbeitsplätze in der Schweiz – Entwicklung und Szenarien bis 2040, im Auftrag des Bundesamtes für Raumentwicklung, Bern: Ecoplan, Link
- License
- GPL >= 2 (Tobias Schoch, vers. 3, July 12, 2020)
- Download
- ipf.R
Robust estimation of the quintile share ratio
- Topic
- Income inequality and poverty measures
- Description
- The quintile share ratio (QSR) of disposable household income is the primary indicator of income inequality in the European Union. As an inequality indicator, it must be sensitive to extreme large observations. However, outliers can have a strong impact on the bias and the variance of the classical estimator which may mislead the interpretation of income inequality. As a remedy, we develop a class of estimators which are robust against outliers.
- Reference
- Hulliger, B. and T. Schoch, 2014. Robust, distribution-free inference for income share ratios under complex sampling, AStA Advances in Statistical Analysis 98, pp. 63-83, Link
- License
- GPL >= 2 (Tobias Schoch and Beat Hulliger, vers. 0.2, 2014)
- Dependencies
- R package robsurvey
- Download
- qsr.R