Predicting Food Consumption to Reduce the Risk of Food Insecurity in Kazakhstan

Abstract

In modern times, the risk of food insecurity is a concern for policymakers at the global and national levels, as the issue of hunger and malnutrition still exists. Food security is vulnerable to any crises. The main goal of this paper is to create a neural-network-based predictive model to forecast food consumption trends in Kazakhstan, aiming to reduce the risk of food insecurity. The initial phase of this study involved identifying socioeconomic factors that significantly influence food consumption behaviors in Kazakhstan. Principal component analysis was used to identify key variables, which became the basis for modelling artificial neural networks. It was revealed that the poverty rate, GDP per capita, and food price index are pivotal determinants of food consumption in Kazakhstan. Two models were prepared: to predict food consumption on a national scale per capita per month, and to predict the percentage distribution of various food categories. The prediction of the percentage distribution of various food categories in Kazakhstan demonstrates the positive modelling quality indicators and strengthens the assumption that network modelling can be used. Predictions for total food consumption over the next three years indicate declining metrics, raising concerns about the potential food insecurity risk in Kazakhstan.