Hunger Hotspots Prediction: Evaluating the Effectiveness of Data-Driven Prediction for Early Crisis Identification

Bhonde, Dnyanesh Mohan (2025) Hunger Hotspots Prediction: Evaluating the Effectiveness of Data-Driven Prediction for Early Crisis Identification. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

Food insecurity continues to challenge global development, mainly in the regions vulnerable to conflict, climate shocks, and economic instability. Traditional early warning system, such as ipc and fewsnet, often struggle with accuracy, and slow response time. This research examine the effectiveness of the data-driven prediction models for early identification of hunger hotspots using Zimbabwe’s publicly available hunger data. A hybrid machine framework implemented using the five different model such as XG-Boost, Randomforest, decision tree, logistic regression, and LTSM(long short-term memory) along with the stack ensemble model technique. Data preprocessing, temporal feature engineering and transformation was used to enhance the data performance. model performance was evaluated using the standard metrics such as MSE, RSME, and R2 score. The hybrid model shows high predictive accuracy of R2 of 0.85, performing well then the individual models and offering improved robustness in identifying the food insecurity trends. XG-Boost and Decision Tree models also perform well with R2 0.87, while LSTM(Long Short-Term Memory) and Linear regression lagged due to linear limitations. The results show that combining the different models can make early warning for food crisis more accurate. Even though there are some challenges like missing data or differences between the regions, the method used in this study can help NGOs and aid organisations to plan better and reduce the hunger crisis.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Haque, Rejwanul UNSPECIFIED
Subjects:	H Social Sciences > HC Economic History and Conditions > Development Economics Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	25 Aug 2026 11:34
Last Modified:	25 Aug 2026 11:34
URI:	https://norma.ncirl.ie/id/eprint/9623

Actions (login required)

View Item