Combining open data and machine learning to predict food security in Ethiopia

Wesley van der Heijden, Marc van den Homberg, Martijn Marijnis, Marijke de Graaff, Hennie Daniels

Research output: Contribution to conferencePaperOther research output

Abstract

Food security is commonly measured by means of surveys, requiring substantial time and budget. Open data can possibly serve as a cost-effective alternative to predict food security. In this paper a method is proposed that uses open data related to food insecurity drivers to predict food security in Ethiopia at the subnational level. The method is based on an ordinal classification approach with a random forest as underlying algorithm. The model turned out to have an accuracy of approximately 90%. Although using an ordinal approach increases performance, a negative side-effect is that the model struggled to predict records with the label ‘stressed’ as a target. The basis of this effect lays in how probabilities for classes ranked in the middle are calculated. Further research on adding open data sources on other drivers and on finetuning hyperparameters in the modelling is advised before implementing machine learning to predict food security.
Original languageEnglish
Publication statusPublished - Jul 2018
Event2018 International Tech4Dev Conference: UNESCO Chair in Technologies for Development: Voices of the Global South - Lausanne, Switzerland
Duration: 27 Jun 201829 Jun 2018

Conference

Conference2018 International Tech4Dev Conference
CountrySwitzerland
CityLausanne
Period27/06/1829/06/18

Fingerprint

Learning systems
Labels
Costs

Cite this

van der Heijden, W., van den Homberg, M., Marijnis, M., de Graaff, M., & Daniels, H. (2018). Combining open data and machine learning to predict food security in Ethiopia. Paper presented at 2018 International Tech4Dev Conference, Lausanne, Switzerland.
van der Heijden, Wesley ; van den Homberg, Marc ; Marijnis, Martijn ; de Graaff, Marijke ; Daniels, Hennie. / Combining open data and machine learning to predict food security in Ethiopia. Paper presented at 2018 International Tech4Dev Conference, Lausanne, Switzerland.
@conference{63b3d655940a47bdbd62f623af9fbe9e,
title = "Combining open data and machine learning to predict food security in Ethiopia",
abstract = "Food security is commonly measured by means of surveys, requiring substantial time and budget. Open data can possibly serve as a cost-effective alternative to predict food security. In this paper a method is proposed that uses open data related to food insecurity drivers to predict food security in Ethiopia at the subnational level. The method is based on an ordinal classification approach with a random forest as underlying algorithm. The model turned out to have an accuracy of approximately 90{\%}. Although using an ordinal approach increases performance, a negative side-effect is that the model struggled to predict records with the label ‘stressed’ as a target. The basis of this effect lays in how probabilities for classes ranked in the middle are calculated. Further research on adding open data sources on other drivers and on finetuning hyperparameters in the modelling is advised before implementing machine learning to predict food security.",
author = "{van der Heijden}, Wesley and {van den Homberg}, Marc and Martijn Marijnis and {de Graaff}, Marijke and Hennie Daniels",
year = "2018",
month = "7",
language = "English",
note = "2018 International Tech4Dev Conference : UNESCO Chair in Technologies for Development: Voices of the Global South ; Conference date: 27-06-2018 Through 29-06-2018",

}

van der Heijden, W, van den Homberg, M, Marijnis, M, de Graaff, M & Daniels, H 2018, 'Combining open data and machine learning to predict food security in Ethiopia' Paper presented at 2018 International Tech4Dev Conference, Lausanne, Switzerland, 27/06/18 - 29/06/18, .

Combining open data and machine learning to predict food security in Ethiopia. / van der Heijden, Wesley; van den Homberg, Marc; Marijnis, Martijn; de Graaff, Marijke; Daniels, Hennie.

2018. Paper presented at 2018 International Tech4Dev Conference, Lausanne, Switzerland.

Research output: Contribution to conferencePaperOther research output

TY - CONF

T1 - Combining open data and machine learning to predict food security in Ethiopia

AU - van der Heijden, Wesley

AU - van den Homberg, Marc

AU - Marijnis, Martijn

AU - de Graaff, Marijke

AU - Daniels, Hennie

PY - 2018/7

Y1 - 2018/7

N2 - Food security is commonly measured by means of surveys, requiring substantial time and budget. Open data can possibly serve as a cost-effective alternative to predict food security. In this paper a method is proposed that uses open data related to food insecurity drivers to predict food security in Ethiopia at the subnational level. The method is based on an ordinal classification approach with a random forest as underlying algorithm. The model turned out to have an accuracy of approximately 90%. Although using an ordinal approach increases performance, a negative side-effect is that the model struggled to predict records with the label ‘stressed’ as a target. The basis of this effect lays in how probabilities for classes ranked in the middle are calculated. Further research on adding open data sources on other drivers and on finetuning hyperparameters in the modelling is advised before implementing machine learning to predict food security.

AB - Food security is commonly measured by means of surveys, requiring substantial time and budget. Open data can possibly serve as a cost-effective alternative to predict food security. In this paper a method is proposed that uses open data related to food insecurity drivers to predict food security in Ethiopia at the subnational level. The method is based on an ordinal classification approach with a random forest as underlying algorithm. The model turned out to have an accuracy of approximately 90%. Although using an ordinal approach increases performance, a negative side-effect is that the model struggled to predict records with the label ‘stressed’ as a target. The basis of this effect lays in how probabilities for classes ranked in the middle are calculated. Further research on adding open data sources on other drivers and on finetuning hyperparameters in the modelling is advised before implementing machine learning to predict food security.

M3 - Paper

ER -

van der Heijden W, van den Homberg M, Marijnis M, de Graaff M, Daniels H. Combining open data and machine learning to predict food security in Ethiopia. 2018. Paper presented at 2018 International Tech4Dev Conference, Lausanne, Switzerland.