Data-driven poverty rate prediction in tunisia using a fuzzy-exponential framework

Abstract

This study introduces a hybrid approach to poverty forecasting by combining fuzzy logic with traditional time series models. Using Tunisia’s poverty rate data from 2000 to 2024, the research applies exponential smoothing as a baseline and integrates fuzzy membership functions to better capture uncertainty. The proposed model improves prediction accuracy, offering a more flexible and adaptive tool for economic policy planning. Results highlight the effectiveness of fuzzy-based forecasting in addressing the limitations of conventional methods in dynamic socioeconomic contexts. Moreover, the proposed methodology is adaptable and can be tailored to reflect the specific characteristics of different countries and contexts, ensuring relevance and effectiveness across diverse settings.