Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Titre : The Future of Food Supply in the Middle East : Case Studies of Iran, Turkey, and Iraq

Auteur : Nosratabadi, Saeed

Université de soutenance : Hungarian University of Agriculture and Life Sciences (formerly Szent Istvan University)

Grade : Ph.D. Dissertation 2021

Sommaire partiel
On the one hand, drought and climate change have reduced agricultural production, which in turn has reduced food supply. On the other hand, the population of different countries, especially in the Middle East, is expected to increase, which will increase the demand for food. This decrease in food production and the increase in food demand create a gap that imposes food insecurity on these countries. Various solutions have been proposed to bridge this gap and address food insecurity. Various stakeholders from various industries, such as the food and agriculture industry, research institutes, and policymakers, have come up with solutions to address the growing demand for food. In fact, the goal is not just to provide food, but to increase supply for the growing demand for a healthy population. Before implementing any program to achieve food security, knowing the amount of food products produced in the country can be a basis for planning and designing solutions to achieve food security. In other words, predicting the future of a country’s food production can provide macro-level decision makers with a picture to use to plan for tackling food insecurity and its consequences. There are many methods for predicting time series data in the literature, but due to the high predictive performance of machine learning models and deep learning models, the use of these models has become very common. Therefore, the present study is conducted to find a proper machine learning model to predict food production (i.e., agricultural production and livestock production) produced locally in Iran, Iraq, and Turkey. In this study, the predictive performance of two machine learning models, namely MLP and ANFIS, was compared on time series data of agricultural products and livestock products taken from FAOSATAT (i.e., FAO database). For this purpose, first the mentioned models were trained by 70% of the data and then their predictive performance was tested with the rest 30% of the data. Initially, the findings showed that the ANFIS model had a higher predictive performance than the MLP model in both predicting agricultural production and livestock production in all three countries studied. Therefore, this model was used to predict agricultural and livestock production for Iran, Turkey, and Iraq for the next ten years. The results of the forecasts disclosed that agricultural production in all three countries of Iran, Turkey and Iraq is expected to increase remarkably in 2030 compared to 2020. However, the results of livestock production forecasts in these three countries revealed that livestock production is expected to increase only in Iran and Iraq, and livestock production in Turkey will decrease in 2030 compared to 2020.

Présentation

Version intégrale ( 1 Mb)