Simulations conducted to compare different estimators for problems containing measuremet errors and non-probability samples.

Joint estimation - based on an article 'Inference for Regression with Variables Generated by AI or Machine Learning'. Compares described in the article joint estimation method implemented using R package optimx with naive one for multilabel classification with measurement error.

Bias corrected estimator - compares efficiency of the bca and bcm estimators from the article above with the naive estimator that does not correct measurement error.

Upgraded bca estimator - compares bca estimator from the article above with its improved version that is valid even in case of non-balanced occuring of measurement error. Both estimators are compared to a naive estimator that does not corrects measurement error at all and precise one that uses precisely measured data.

Combining probability and nonprobability samplings - repeating simulation from an article 'Combining non-probability and probability survey samples through mass imputation' but with data containing measurement error.

GLM bca estimators - measures efficiency of the bca estimator adapted to work in generalised linear models