HYBRID METHOD IMPLEMENTATION: FUZZY DECISION TREE IN THE CLASSIFICATION OF GENDER INEQUALITY

The classification of continuous data using the C4.5 decision tree algorithm requires prior discretization based on the calculation of cut points, a process that can be time-consuming and potentially introduce bias. These limitations may negatively impact both the computational efficiency and the classification accuracy of the decision tree model. This study proposes a hybrid method that integrates fuzzy logic with decision tree techniques in the classification process of continuous data types. Fuzzy logic is utilized to manage uncertainty in data variables and enhance flexibility in processing continuous values, while the decision tree plays a role in providing a structured and rule-based framework for decision-making. This proposed method is applied to gender inequality data, encompassing aspects of reproductive health, education and empowerment, and employment across 166 countries worldwide. The results demonstrate that the fuzzy decision tree method, which was constructed using the C4.5 algorithm, achieved a classification accuracy of 91%, while the C4.5 decision tree method without fuzzy only achieved a classification accuracy of 77%. The fuzzy decision tree method successfully improved the classification accuracy by 14%. Additionally, the fuzzy decision tree exhibited more stable and balanced performance in classifying data into four target categories. Therefore, this approach offers an effective and comprehensive alternative for classifying gender inequality, with the potential to support data-driven and targeted policy-making.