Expert System for Poverty Level Assessment Using the Certainty Factor Method

Abstract

Poverty is a multidimensional socio-economic problem that remains a major challenge in Indonesia’s development process. The complexity of poverty is influenced by various interconnected factors, including economic conditions, education, human development, access to basic services, and employment opportunities. Accurate identification of poverty levels is therefore essential to support effective policy formulation and targeted poverty alleviation programs. This study aims to develop an expert system for assessing poverty levels in regencies and cities throughout Indonesia using the Certainty Factor (CF) method. The CF approach was selected because it is capable of representing expert knowledge and handling uncertainty in decision making processes. The proposed system utilizes seven socio-economic indicators that are widely recognized as determinants of poverty, namely the percentage of the poor population, mean years of schooling, per capita expenditure, the Human Development Index (HDI), access to decent sanitation, access to safe drinking water, and the open unemployment rate. Expert certainty values were established through an extensive literature review and incorporated into the system as knowledge-based rules. These values were then combined with user certainty values derived from observational data collected from 514 regencies and cities across Indonesia. The integration of expert knowledge and empirical data enables the system to generate a
poverty-level assessment that is both transparent and measurable. The performance of the developed expert system was evaluated by comparing the classification results with actual poverty data from the
514 regencies and cities. The evaluation results demonstrate that the system achieved an accuracy of 90.66%, precision of 58.75%, recall of 75.81%, and an F1-score of 66.20%. These findings indicate that
the proposed approach is effective in identifying regions with varying poverty levels and has strong potential as a decision-support tool. The developed system can assist policymakers and stakeholders in
prioritizing interventions, allocating resources more efficiently, and supporting evidence-based poverty reduction strategies across Indonesia.