Package 'saeczi'

Title: Small Area Estimation for Continuous Zero Inflated Data
Description: Provides functionality to fit a zero-inflated estimator for small area estimation. This estimator is a combines a linear mixed effects regression model and a logistic mixed effects regression model via a two-stage modeling approach. The estimator's mean squared error is estimated via a parametric bootstrap method. Chandra and others (2012, <doi:10.1080/03610918.2011.598991>) introduce and describe this estimator and mean squared error estimator. White and others (2024+, <doi:10.48550/arXiv.2402.03263>) describe the applicability of this estimator to estimation of forest attributes and further assess the estimator's properties.
Authors: Josh Yamamoto [aut, cre], Dinan Elsyad [aut], Grayson White [aut], Julian Schmitt [aut], Niels Korsgaard [aut], Kelly McConville [aut], Kate Hu [aut]
Maintainer: Josh Yamamoto <[email protected]>
License: MIT + file LICENSE
Version: 0.2.0.9000
Built: 2025-03-09 05:57:28 UTC
Source: https://github.com/harvard-ufds/saeczi

Help Index

FIA Population Level Auxiliary Data for Oregon

Description

FIA Population Level Auxiliary Data for Oregon

Usage

pop

Format

An object of class data.frame with 10060 rows and 10 columns.

Fit a zero-inflation estimator.

Description

Fit a zero-inflation estimator.

Usage

saeczi(
  samp_dat,
  pop_dat,
  lin_formula,
  log_formula = lin_formula,
  domain_level,
  B = 100L,
  mse_est = FALSE,
  estimand = "means",
  parallel = FALSE,
  transform_fun = NULL,
  inv_transform_fun = NULL
)

Arguments

samp_dat

A data.frame with domains, auxiliary variables, and the response variable of a sample
pop_dat

A data.frame with domains and auxiliary variables of a population.
lin_formula

Formula. Specification of the response and fixed effects of the linear regression model
log_formula

Formula. Specification of the response and fixed effects of the logistic regression model
domain_level

String. The column name in samp_dat and pop_dat that encodes the domain level
B

Integer. The number of bootstraps to be used in MSE estimation.
mse_est

Logical. Whether or not MSE estimation should happen.
estimand

String. Whether the estimates should be 'totals' or 'means'.
parallel

Logical. Should the MSE estimation be computed in parallel.
transform_fun

Function. Function to be applied to the response variable prior to modeling.
inv_transform_fun

Function. Inverse of transform_fun. Required if transform_fun is specified.

Value

An object of class 'zi_mod' with defined 'print()' and 'summary()' methods. The object is structured like a list and contains the following elements:

* call: The original function call

* res: A data.frame containing the estimates and mse estimates

* lin_mod: The modeling object used to fit the original linear model

* log_mod: The modeling object used to fit the original logistic model

Examples

data(pop)
data(samp)

lin_formula <- DRYBIO_AG_TPA_live_ADJ ~ tcc16 + elev

result <- saeczi(samp_dat = samp,
                 pop_dat = pop,
                 lin_formula = lin_formula,
                 log_formula = lin_formula,
                 domain_level = "COUNTYFIPS",
                 mse_est = FALSE)

FIA sample data for Oregon

Description

FIA sample data for Oregon

Usage

samp

Format

An object of class tbl_df (inherits from tbl, data.frame) with 1494 rows and 11 columns.