Is Random Forest a Superior Methodology for Predicting Poverty? : An Empirical Assessment

Random forest is in many fields of research a common method for data driven predictions. Within economics and prediction of poverty, random forest is rarely used. Comparing out-of-sample predictions in surveys for same year in six countries shows t...

Full description

Bibliographic Details
Main Authors:	Sohnesen, Thomas Pave, Stender, Niels
Format:	Working Paper
Language:	English en_US
Published:	World Bank, Washington, DC 2016
Subjects:	PREDICTIONS POOR HOUSEHOLD CONSUMPTION EXPENDITURES HOUSEHOLD SIZE HOUSEHOLD SURVEY AGRICULTURAL GROWTH CONSUMPTION POVERTY REDUCTION IMPACT ON POVERTY POVERTY RATES ERRORS FARMER POVERTY RATE FOOD CONSUMPTION INCOME LINEAR REGRESSION POVERTY ESTIMATES ALGORITHMS HOUSEHOLD SURVEYS PROGRAMS CONSUMPTION DATA HOUSING AGRICULTURAL PRACTICES IMPACTS NATIONAL POVERTY SAMPLES RURAL VARIABLES MEASUREMENT COUNTING HOUSEHOLD BUDGET CONSUMPTION AGGREGATE QUALITY SURVEYS SOCIAL ASSISTANCE MEASURES INSTRUMENTS TARGETING RANDOM SAMPLES CONSUMPTION EXPENDITURE RURAL AREAS CROSS‐SECTION DATA WELFARE MEASURES WELFARE INDICATORS PANEL DATA SETS REGIONS STATISTICS EVALUATION SIGNIFICANCE LEVEL POOR HOUSEHOLDS SAMPLING POVERTY HOUSEHOLD HEAD HOUSEHOLD CONSUMPTION ECONOMETRICS STANDARD ERRORS POVERTY STATUS POOR PREDICTION POVERTY ASSESSMENT LEARNING INDICATORS RESEARCH CONSUMPTION POVERTY OUTCOMES SOCIAL INDICATORS MISSING OBSERVATIONS INEQUALITY
Online Access:	http://documents.worldbank.org/curated/en/2016/03/26089791/random-forest-superior-methodology-predicting-poverty-empirical-assessment http://hdl.handle.net/10986/24154

id	okr-10986-24154
recordtype	oai_dc
spelling	okr-10986-241542021-04-23T14:04:19Z Is Random Forest a Superior Methodology for Predicting Poverty? : An Empirical Assessment Sohnesen, Thomas Pave Stender, Niels PREDICTIONS POOR HOUSEHOLD CONSUMPTION EXPENDITURES HOUSEHOLD SIZE HOUSEHOLD SURVEY AGRICULTURAL GROWTH CONSUMPTION POVERTY REDUCTION IMPACT ON POVERTY POVERTY RATES ERRORS FARMER POVERTY RATE FOOD CONSUMPTION INCOME LINEAR REGRESSION POVERTY RATES POVERTY ESTIMATES ALGORITHMS HOUSEHOLD SURVEYS PROGRAMS CONSUMPTION DATA HOUSEHOLD SIZE HOUSING POVERTY ESTIMATES AGRICULTURAL PRACTICES IMPACTS NATIONAL POVERTY SAMPLES RURAL VARIABLES MEASUREMENT COUNTING HOUSEHOLD BUDGET CONSUMPTION AGGREGATE QUALITY SURVEYS SOCIAL ASSISTANCE MEASURES INSTRUMENTS POVERTY REDUCTION TARGETING RANDOM SAMPLES AGRICULTURAL PRACTICES CONSUMPTION EXPENDITURE RURAL AREAS CROSS‐SECTION DATA WELFARE MEASURES CROSS‐SECTION DATA WELFARE INDICATORS PANEL DATA SETS SOCIAL ASSISTANCE REGIONS STATISTICS EVALUATION SIGNIFICANCE LEVEL POOR HOUSEHOLDS SAMPLING RURAL AREAS POVERTY POOR HOUSEHOLD HOUSEHOLD HEAD PANEL DATA SETS CONSUMPTION EXPENDITURES SIGNIFICANCE LEVEL NATIONAL POVERTY HOUSEHOLD CONSUMPTION ECONOMETRICS STANDARD ERRORS CONSUMPTION DATA POVERTY STATUS POVERTY RATE POOR PREDICTION POVERTY ASSESSMENT CONSUMPTION EXPENDITURE HOUSEHOLD SURVEYS LEARNING INDICATORS RESEARCH CONSUMPTION POVERTY WELFARE INDICATORS OUTCOMES SOCIAL INDICATORS POVERTY STATUS LINEAR REGRESSION MISSING OBSERVATIONS INEQUALITY POOR HOUSEHOLDS Random forest is in many fields of research a common method for data driven predictions. Within economics and prediction of poverty, random forest is rarely used. Comparing out-of-sample predictions in surveys for same year in six countries shows that random forest is often more accurate than current common practice (multiple imputations with variables selected by stepwise and Lasso), suggesting that this method could contribute to better poverty predictions. However, none of the methods consistently provides accurate predictions of poverty over time, highlighting that technical model fitting by any method within a single year is not always, by itself, sufficient for accurate predictions of poverty over time. 2016-04-26T16:49:08Z 2016-04-26T16:49:08Z 2016-03 Working Paper http://documents.worldbank.org/curated/en/2016/03/26089791/random-forest-superior-methodology-predicting-poverty-empirical-assessment http://hdl.handle.net/10986/24154 English en_US Policy Research Working Paper;No. 7612 CC BY 3.0 IGO http://creativecommons.org/licenses/by/3.0/igo/ World Bank World Bank, Washington, DC Publications & Research Publications & Research :: Policy Research Working Paper
repository_type	Digital Repository
institution_category	Foreign Institution
institution	Digital Repositories
building	World Bank Open Knowledge Repository
collection	World Bank
language	English en_US
topic	PREDICTIONS POOR HOUSEHOLD CONSUMPTION EXPENDITURES HOUSEHOLD SIZE HOUSEHOLD SURVEY AGRICULTURAL GROWTH CONSUMPTION POVERTY REDUCTION IMPACT ON POVERTY POVERTY RATES ERRORS FARMER POVERTY RATE FOOD CONSUMPTION INCOME LINEAR REGRESSION POVERTY RATES POVERTY ESTIMATES ALGORITHMS HOUSEHOLD SURVEYS PROGRAMS CONSUMPTION DATA HOUSEHOLD SIZE HOUSING POVERTY ESTIMATES AGRICULTURAL PRACTICES IMPACTS NATIONAL POVERTY SAMPLES RURAL VARIABLES MEASUREMENT COUNTING HOUSEHOLD BUDGET CONSUMPTION AGGREGATE QUALITY SURVEYS SOCIAL ASSISTANCE MEASURES INSTRUMENTS POVERTY REDUCTION TARGETING RANDOM SAMPLES AGRICULTURAL PRACTICES CONSUMPTION EXPENDITURE RURAL AREAS CROSS‐SECTION DATA WELFARE MEASURES CROSS‐SECTION DATA WELFARE INDICATORS PANEL DATA SETS SOCIAL ASSISTANCE REGIONS STATISTICS EVALUATION SIGNIFICANCE LEVEL POOR HOUSEHOLDS SAMPLING RURAL AREAS POVERTY POOR HOUSEHOLD HOUSEHOLD HEAD PANEL DATA SETS CONSUMPTION EXPENDITURES SIGNIFICANCE LEVEL NATIONAL POVERTY HOUSEHOLD CONSUMPTION ECONOMETRICS STANDARD ERRORS CONSUMPTION DATA POVERTY STATUS POVERTY RATE POOR PREDICTION POVERTY ASSESSMENT CONSUMPTION EXPENDITURE HOUSEHOLD SURVEYS LEARNING INDICATORS RESEARCH CONSUMPTION POVERTY WELFARE INDICATORS OUTCOMES SOCIAL INDICATORS POVERTY STATUS LINEAR REGRESSION MISSING OBSERVATIONS INEQUALITY POOR HOUSEHOLDS
spellingShingle	PREDICTIONS POOR HOUSEHOLD CONSUMPTION EXPENDITURES HOUSEHOLD SIZE HOUSEHOLD SURVEY AGRICULTURAL GROWTH CONSUMPTION POVERTY REDUCTION IMPACT ON POVERTY POVERTY RATES ERRORS FARMER POVERTY RATE FOOD CONSUMPTION INCOME LINEAR REGRESSION POVERTY RATES POVERTY ESTIMATES ALGORITHMS HOUSEHOLD SURVEYS PROGRAMS CONSUMPTION DATA HOUSEHOLD SIZE HOUSING POVERTY ESTIMATES AGRICULTURAL PRACTICES IMPACTS NATIONAL POVERTY SAMPLES RURAL VARIABLES MEASUREMENT COUNTING HOUSEHOLD BUDGET CONSUMPTION AGGREGATE QUALITY SURVEYS SOCIAL ASSISTANCE MEASURES INSTRUMENTS POVERTY REDUCTION TARGETING RANDOM SAMPLES AGRICULTURAL PRACTICES CONSUMPTION EXPENDITURE RURAL AREAS CROSS‐SECTION DATA WELFARE MEASURES CROSS‐SECTION DATA WELFARE INDICATORS PANEL DATA SETS SOCIAL ASSISTANCE REGIONS STATISTICS EVALUATION SIGNIFICANCE LEVEL POOR HOUSEHOLDS SAMPLING RURAL AREAS POVERTY POOR HOUSEHOLD HOUSEHOLD HEAD PANEL DATA SETS CONSUMPTION EXPENDITURES SIGNIFICANCE LEVEL NATIONAL POVERTY HOUSEHOLD CONSUMPTION ECONOMETRICS STANDARD ERRORS CONSUMPTION DATA POVERTY STATUS POVERTY RATE POOR PREDICTION POVERTY ASSESSMENT CONSUMPTION EXPENDITURE HOUSEHOLD SURVEYS LEARNING INDICATORS RESEARCH CONSUMPTION POVERTY WELFARE INDICATORS OUTCOMES SOCIAL INDICATORS POVERTY STATUS LINEAR REGRESSION MISSING OBSERVATIONS INEQUALITY POOR HOUSEHOLDS Sohnesen, Thomas Pave Stender, Niels Is Random Forest a Superior Methodology for Predicting Poverty? : An Empirical Assessment
relation	Policy Research Working Paper;No. 7612
description	Random forest is in many fields of research a common method for data driven predictions. Within economics and prediction of poverty, random forest is rarely used. Comparing out-of-sample predictions in surveys for same year in six countries shows that random forest is often more accurate than current common practice (multiple imputations with variables selected by stepwise and Lasso), suggesting that this method could contribute to better poverty predictions. However, none of the methods consistently provides accurate predictions of poverty over time, highlighting that technical model fitting by any method within a single year is not always, by itself, sufficient for accurate predictions of poverty over time.
format	Working Paper
author	Sohnesen, Thomas Pave Stender, Niels
author_facet	Sohnesen, Thomas Pave Stender, Niels
author_sort	Sohnesen, Thomas Pave
title	Is Random Forest a Superior Methodology for Predicting Poverty? : An Empirical Assessment
title_short	Is Random Forest a Superior Methodology for Predicting Poverty? : An Empirical Assessment
title_full	Is Random Forest a Superior Methodology for Predicting Poverty? : An Empirical Assessment
title_fullStr	Is Random Forest a Superior Methodology for Predicting Poverty? : An Empirical Assessment
title_full_unstemmed	Is Random Forest a Superior Methodology for Predicting Poverty? : An Empirical Assessment
title_sort	is random forest a superior methodology for predicting poverty? : an empirical assessment
publisher	World Bank, Washington, DC
publishDate	2016
url	http://documents.worldbank.org/curated/en/2016/03/26089791/random-forest-superior-methodology-predicting-poverty-empirical-assessment http://hdl.handle.net/10986/24154
_version_	1764455791054553088

Is Random Forest a Superior Methodology for Predicting Poverty? : An Empirical Assessment

Similar Items