Skip to content

NMAR 0.1.0

Choose a tag to compare

View all tags
@Iwaniukooo11 Iwaniukooo11 released this 19 Dec 19:21
v0.1.0
2034668
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

NMAR_0.1.0.tar.gz

Initial CRAN Release

  • First release of the NMAR package for estimating mean under nonignorable nonresponse (NMAR) models in sample surveys.

Methods

  • Empirical Likelihood (EL): Added el_engine() implementing the estimator of Qin, Leung, and Shao (2002). This method uses empirical likelihood weights satisfying response mechanism equations and auxiliary moment constraints.
  • Exponential Tilting (Parametric & Nonparametric): Included robust implementations for both microdata (exptilt_engine) and aggregated contingency tables (exptilt_nonparam_engine) based on Riddles, Kim, and Im (2016).

Key Features

  • Unified API: All estimators are accessible via a single, consistent nmar() interface supporting standard formula syntax (e.g., Y ~ X | Z).
  • Complex Survey Support: Seamless integration with the survey package. nmar() accepts survey.design objects, automatically handling weights and stratification.
  • Variance Estimation: Robust bootstrapping (S3) implementation for standard errors and confidence intervals across all engines.
  • Diagnostics: Rich return objects including convergence statistics, Jacobian condition numbers, and detailed weight summaries.

Major Changes

  • Refactored Architecture: The exptilt and el engines share a unified structural design, ensuring consistent behavior for controls, standardization, and error handling.
  • Standardization: Added standardize = TRUE argument to engines to improve numerical stability during optimization.
Assets 3
Loading