Search results

1

Estimation of Parameters for Small Areas Using Hierarchical Bayes Method in the Case of Known Model Hyperparameters

100%

Kubacki J.

Statistics in Transition new series

|

2012

|

vol. 13

|

issue 2

261-278

EN

In the paper the method of parameters estimation using hierarchical Bayes (HB) method in the case of known model hyperparameters for a priori conditionals was presented. This approach has some advantage in comparison with subjective model parameters selection because of more simulation stability and allows obtaining estimates that has more regular distribution. As an example the data about average per capita income from Polish Household Budget Survey for counties (NUTS4) and auxiliary variables from Polish Tax Register (POLTAX) were used. The computation was done using WinBUGS software and R-project environment with R2WinBUGS package, which control the simulations in WinBUGS, and coda package, which allows performing the analysis of simulation results. In the paper sample code in R-project that can be used as a pattern for further similar applications was also presented. The efficiency of hierarchical Bayes estimation with other small area methods was compared. Such comparison was done for HB and EBLUP techniques, for which some consistency related to the precision of estimates obtained using both techniques was achieved.

2

Some Applications of Panel Data Models in Small Area Estimation

80%

Nekrašaitė-Liegė V.

Statistics in Transition new series

|

2011

|

vol. 12

|

issue 2

265-280

EN

This study uses a real population from Statistics Lithuania to investigate the performance of different types of estimation strategies. The estimation strategy is a combination of sampling design and estimation design. The sampling designs include equal probability design (SRS) and unequal probability designs (stratified SRS and model-based sampling designs). Design-based direct Horvitz-Thompson, indirect model-assisted GREG estimator and indirect model-based estimator are used to estimate the totals in small area estimation. The underlying panel-type models (linear fixed-effects type or linear random-effects type) are examined in both stages of estimation strategies: sample design and construction of estimators.

3

Small Area Estimation Under a Mixture Model

70%

Chandra H., Bathla H., Sud U. C.

Statistics in Transition new series

|

2010

|

vol. 11

|

issue 3

76-89

EN

Small area estimation (SAE) under a linear mixed model may not be efficient if data contain substantial proportion of zeros than would be expected under standard model assumptions (hereafter zero-inflated data). We discuss the SAE for zero-inflated data under a mixture model (Fletcher et al., 2005 and Karlberg, 2000) that account for excess zeros in the data. Our results from simulation studies show that mixture model based approach for SAE works well and produces an efficient set of small area estimates. An application to real survey data from the National Sample Survey Organisation of India demonstrates the satisfactory performance of the approach.

Refine search results

3 Statistics in Transition new series

1 Bathla H.

1 Chandra H.

1 Kubacki J.

1 Nekrašaitė-Liegė V.

1 Sud U. C.

1 2012

1 2011

1 2010

Estimation of Parameters for Small Areas Using Hierarchical Bayes Method in the Case of Known Model Hyperparameters

Some Applications of Panel Data Models in Small Area Estimation

Small Area Estimation Under a Mixture Model