saePy 0.2.0

Small Area Estimation Python package

saePy

saePy is a Python package for Small Area Estimation (SAE) implementing the EBLUP Fay-Herriot (FH), Nested Error Regression (BHF), and spatial Fay-Herriot (SFH) models. It also provides functions for estimating Mean Squared Error (MSE) and parametric bootstrap MSE for these models.

Features

• EBLUP Fay-Herriot (FH): Area-level small area estimation.
• EBLUP Nested Error Regression (BHF): Unit-level small area estimation.
• EBLUP Spatial Fay-Herriot (SFH): Area-level model with spatial correlation.
• MSE Estimation: Functions to estimate the mean squared error for FH and SFH models.
• Parametric Bootstrap MSE for BHF: Bootstrap-based MSE estimation for the BHF model.
• Example Datasets: Several example datasets are included in the saePy/ directory.

Installation

Make sure you have the following dependencies installed:

• numpy
• scipy
• pandas
• patsy
• statsmodels

Install the package using:

pip install saePy

or

python setup.py install

Directory Structure

saePy/
    __init__.py
    eblupBHF.py
    eblupFH.py
    eblupSFH.py
    mseFH.py
    mseSFH.py
    pbmseBHF.py
    dataset_1.csv
    grapes.csv
    grapesprox.csv
    unit_level_1.csv

Usage

EBLUP Fay-Herriot

from saePy.eblupFH import eblupFH
import pandas as pd

data = pd.read_csv("saePy/dataset_1.csv")
result = eblupFH(
    formula="y ~ x1 + x2",
    vardir_col="vardir",
    data=data
)
print(result)

EBLUP BHF (Unit Level)

from saePy.eblupBHF import eblupBHF

result = eblupBHF(
    formula="y ~ x1 + x2",
    dom="domain",
    selectdom=[...],           # list of domains to estimate
    meanxpop=...,              # population mean matrix
    popnsize=...,              # population size per domain
    data=data
)
print(result)

MSE Estimation

from saePy.mseFH import mseFH

result = mseFH(
    formula="y ~ x1 + x2",
    vardir_col="vardir",
    data=data
)
print(result)

Parametric Bootstrap MSE for BHF

from saePy.pbmseBHF import pbmseBHF

result = pbmseBHF(
    formula="y ~ x1 + x2",
    dom="domain",
    selectdom=[...],
    meanxpop=...,
    popnsize=...,
    data=data
)
print(result)

License

MIT License

---

Note:

• Ensure your data columns match the function arguments.
• For more details, see the documentation in each module in saePy/.