A Comparison of Machine Learning and Deep Learning Models for Predicting Household Food Security Status

Abstract

ML and DL algorithms are becoming more popular to predict household food security status, which can be used by the governments and policymakers of the country to provide a food supply for the needy in case of emergency. ML models, namely: k-Nearest Neighbor (kNN), Logistic Regression (LR), Support Vector Machine (SVM), Naive Bayes (NB), Multi-Layer Perceptron (MLP) and DL models, namely: Artificial Neural Network (ANN) and Convolutional Neural network (CNN) are investigated to predict household food security status in Household Income, Consumption and Expenditure (HICE) survey data of Ethiopia. The standard evaluation measures such as accuracy, precision, recall, F1-score, Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) are used to evaluate the models' predictive performance, and the experimental results reveal that ANN, a DL model surpassed the ML classifiers with an accuracy of 99.15%