saekernel

Propose an area-level, non-parametric regression estimator based on Nadaraya-Watson kernel on small area mean. Adopt a two-stage estimation approach proposed by Prasad and Rao (1990). MSE estimators are not readily available, so resampling method that called bootstrap is applied. This package are based on the model proposed in Two stage non-parametric approach for small area estimation by Pushpal Mukhopadhyay and Tapabrata Maiti.

Installation

You can install the released version of saekernel from CRAN with:

install.packages("saekernel")

Authors

Wicak Surya Hasani, Azka Ubaidillah

Maintainer

Wicak Surya Hasani 221710052@stis.ac.id

Functions

• saekernel() Produces Small Area Estimation Non-Parametric Based Nadaraya-Watson Kernel
• mse_saekernel() Produces Small Area Estimation Non-Parametric based Nadaraya-Watson Kernel and Bootstrap Mean Squared Error

References

• Mukhopadhyay P, Maiti T. (2004). Two Stage Non-Parametric Approach for Small Areas Estimation. Proceedings of ASA Section on Survey Research Methods: 4058-4065.
• Prasad, N.G.N., and Rao, J.N.K. (1990). “The estimation of the mean squared error of the small area estimators.” Journal of the American statistical association. 85. 163-171.
• Hardle, W. (2002). “Applied non-parametric regression,” Cambridge University Press.
• Indahwati, Sadik K, Nurmasari R. (2008). Pendekatan Metode Pemulusan Kernel Pada Pendugaan Area Kecil. Makalah Semnas Matematika. Universitas Negeri Yogyakarta. Yogyakarta.
• Darsyah, M. Y. (2013). Small Area Estimation terhadap Pengeluaran Per Kapita di Kabupaten Sumenep Dengan Pendekatan Nonparametrik. Jurnal Statistika Volume 1 Nomor 2. Universitas Muhammadiyah Semarang.