Poverty Mapping in the Age of Machine Learning
Recent years have witnessed considerable methodological advances in poverty mapping, much of which has focused on the application of modern machine-learning approaches to remotely sensed data. Poverty maps produced with these methods generally share a common validation procedure, which assesses model performance by comparing subnational machine-learning-based poverty estimates with survey-based, direct estimates. Although unbiased, survey-based estimates at a granular level can be imprecise measures of true poverty rates, meaning that it is unclear whether the validation procedures used in machine-learning approaches are informative of actual model performance. This paper examines the credibility of existing approaches to model validation by constructing a pseudo-census from the Mexican Intercensal Survey of 2015, which is used to conduct several design-based simulation experiments. The findings show that the validation procedure often used for machine-learning approaches can be misleading in terms of model assessment since it yields incorrect information for choosing what may be the best set of estimates across different methods and scenarios. Using alternative validation methods, the paper shows that machine-learning-based estimates can rival traditional, more data intensive poverty mapping approaches. Further, the closest approximation to existing machine-learning approaches, using publicly available geo-referenced data, performs poorly when evaluated against “true” poverty rates and fails to outperform traditional poverty mapping methods in targeting simulations.
Main Authors: | , , |
---|---|
Format: | Working Paper biblioteca |
Language: | English English |
Published: |
World Bank, Washington, DC
2023-05-04
|
Subjects: | SMALL AREA ESTIMATION, POVERTY MAPPING, MACHINE LEARNING, SATELLITE IMAGERY, POVERTY RATE ESTIMATE, POVERTY DATA ANALYSIS, |
Online Access: | http://documents.worldbank.org/curated/en/099759405012313710/IDU0c1878458051a40470808a960c8b70982671b https://openknowledge.worldbank.org/handle/10986/39783 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
dig-okr-1098639783 |
---|---|
record_format |
koha |
spelling |
dig-okr-10986397832024-03-11T19:21:44Z Poverty Mapping in the Age of Machine Learning Corral, Paul Henderson, Heath Segovia, Sandra SMALL AREA ESTIMATION POVERTY MAPPING MACHINE LEARNING SATELLITE IMAGERY POVERTY RATE ESTIMATE POVERTY DATA ANALYSIS Recent years have witnessed considerable methodological advances in poverty mapping, much of which has focused on the application of modern machine-learning approaches to remotely sensed data. Poverty maps produced with these methods generally share a common validation procedure, which assesses model performance by comparing subnational machine-learning-based poverty estimates with survey-based, direct estimates. Although unbiased, survey-based estimates at a granular level can be imprecise measures of true poverty rates, meaning that it is unclear whether the validation procedures used in machine-learning approaches are informative of actual model performance. This paper examines the credibility of existing approaches to model validation by constructing a pseudo-census from the Mexican Intercensal Survey of 2015, which is used to conduct several design-based simulation experiments. The findings show that the validation procedure often used for machine-learning approaches can be misleading in terms of model assessment since it yields incorrect information for choosing what may be the best set of estimates across different methods and scenarios. Using alternative validation methods, the paper shows that machine-learning-based estimates can rival traditional, more data intensive poverty mapping approaches. Further, the closest approximation to existing machine-learning approaches, using publicly available geo-referenced data, performs poorly when evaluated against “true” poverty rates and fails to outperform traditional poverty mapping methods in targeting simulations. 2023-05-04T13:53:48Z 2023-05-04T13:53:48Z 2023-05-04 Working Paper http://documents.worldbank.org/curated/en/099759405012313710/IDU0c1878458051a40470808a960c8b70982671b https://openknowledge.worldbank.org/handle/10986/39783 English en Policy Research Working Papers; 10429 CC BY 3.0 IGO https://creativecommons.org/licenses/by/3.0/igo/ World Bank application/pdf text/plain World Bank, Washington, DC |
institution |
Banco Mundial |
collection |
DSpace |
country |
Estados Unidos |
countrycode |
US |
component |
Bibliográfico |
access |
En linea |
databasecode |
dig-okr |
tag |
biblioteca |
region |
America del Norte |
libraryname |
Biblioteca del Banco Mundial |
language |
English English |
topic |
SMALL AREA ESTIMATION POVERTY MAPPING MACHINE LEARNING SATELLITE IMAGERY POVERTY RATE ESTIMATE POVERTY DATA ANALYSIS SMALL AREA ESTIMATION POVERTY MAPPING MACHINE LEARNING SATELLITE IMAGERY POVERTY RATE ESTIMATE POVERTY DATA ANALYSIS |
spellingShingle |
SMALL AREA ESTIMATION POVERTY MAPPING MACHINE LEARNING SATELLITE IMAGERY POVERTY RATE ESTIMATE POVERTY DATA ANALYSIS SMALL AREA ESTIMATION POVERTY MAPPING MACHINE LEARNING SATELLITE IMAGERY POVERTY RATE ESTIMATE POVERTY DATA ANALYSIS Corral, Paul Henderson, Heath Segovia, Sandra Poverty Mapping in the Age of Machine Learning |
description |
Recent years have witnessed
considerable methodological advances in poverty mapping,
much of which has focused on the application of modern
machine-learning approaches to remotely sensed data. Poverty
maps produced with these methods generally share a common
validation procedure, which assesses model performance by
comparing subnational machine-learning-based poverty
estimates with survey-based, direct estimates. Although
unbiased, survey-based estimates at a granular level can be
imprecise measures of true poverty rates, meaning that it is
unclear whether the validation procedures used in
machine-learning approaches are informative of actual model
performance. This paper examines the credibility of existing
approaches to model validation by constructing a
pseudo-census from the Mexican Intercensal Survey of 2015,
which is used to conduct several design-based simulation
experiments. The findings show that the validation procedure
often used for machine-learning approaches can be misleading
in terms of model assessment since it yields incorrect
information for choosing what may be the best set of
estimates across different methods and scenarios. Using
alternative validation methods, the paper shows that
machine-learning-based estimates can rival traditional, more
data intensive poverty mapping approaches. Further, the
closest approximation to existing machine-learning
approaches, using publicly available geo-referenced data,
performs poorly when evaluated against “true” poverty rates
and fails to outperform traditional poverty mapping methods
in targeting simulations. |
format |
Working Paper |
topic_facet |
SMALL AREA ESTIMATION POVERTY MAPPING MACHINE LEARNING SATELLITE IMAGERY POVERTY RATE ESTIMATE POVERTY DATA ANALYSIS |
author |
Corral, Paul Henderson, Heath Segovia, Sandra |
author_facet |
Corral, Paul Henderson, Heath Segovia, Sandra |
author_sort |
Corral, Paul |
title |
Poverty Mapping in the Age of Machine Learning |
title_short |
Poverty Mapping in the Age of Machine Learning |
title_full |
Poverty Mapping in the Age of Machine Learning |
title_fullStr |
Poverty Mapping in the Age of Machine Learning |
title_full_unstemmed |
Poverty Mapping in the Age of Machine Learning |
title_sort |
poverty mapping in the age of machine learning |
publisher |
World Bank, Washington, DC |
publishDate |
2023-05-04 |
url |
http://documents.worldbank.org/curated/en/099759405012313710/IDU0c1878458051a40470808a960c8b70982671b https://openknowledge.worldbank.org/handle/10986/39783 |
work_keys_str_mv |
AT corralpaul povertymappingintheageofmachinelearning AT hendersonheath povertymappingintheageofmachinelearning AT segoviasandra povertymappingintheageofmachinelearning |
_version_ |
1794796885257486336 |