EVALUATION OF SOCIAL ASSISTANCE RECIPIENT ELIGIBILITY BASED ON INCOME AND NUMBER OF DEPENDENTS USING THE FUZZY MAMDANI APPROACH

Abstract

This study aims to evaluate the eligibility of social assistance recipients by considering income and family dependents using the Mamdani fuzzy logic approach. A key issue in the current selection process of social assistance recipients lies in the limitations of binary classification methods, which are rigid and incapable of fully representing the diversity of socioeconomic conditions in society. By applying the Mamdani fuzzy method, this research incorporates linguistic variables such as "low income" and "large number of dependents" into a more adaptive and continuous assessment system.The study employs primary data collected from 20 respondents, with income ranging from Rp100,000 to Rp900,000 and family dependents ranging from 1 to 6 individuals. The developed fuzzy model uses three membership functions for each input and output variable to enhance the system’s sensitivity to data variations. Fuzzy inference is conducted based on Mamdani rules, while defuzzification utilizes the centroid method to produce eligibility scores on a continuous scale of 0 to 100. The findings demonstrate that the Mamdani fuzzy approach yields a more equitable, transparent, and realistic classification of eligibility for social assistance.As a practical contribution, this study also provides a MATLAB script implementation that can be easily adapted and applied by local governments or social organizations in data-driven selection processes for social assistance recipients.