<![CDATA[The issue of delayed graduation is often encountered in various universities, including in the Statistics Study Program at Universitas Negeri Gorontalo, for graduates between 2018 and 2023. Among them, 162 students (76.5%) experienced delayed graduation, and 5 students (2.35%) dropped out. This delay in graduation is caused by various factors, necessitating a classification method capable of identifying the most dominant factors. The classification method used in this research is Improved Chi-Square Automatic Interaction Detection (I-CHAID) with the Synthetic Minority Oversampling Technique (SMOTE) approach. SMOTE is employed to address imbalanced data. Based on the I-CHAID classification tree with the SMOTE approach, the significant factors influencing the duration of study completion are the GPA in the fifth semester (67.2%) and the mentoring method (87.5%). As for the classification performance from the 40% testing data, the accuracy achieved was 40.6%, meaning that out of 32 samples, 13 were correctly classified. The sensitivity value was 6.25%, indicating the success rate of classifying data for students who graduated on time. The specificity value was 75%, showing the success rate in classifying data for students who did not graduate on time. The precision value was 20%, reflecting the accuracy of predicting students who actually graduated on time, and the F-measure was 9.52%, indicating the balance between precision and sensitivity.]]>
Copyrights © 2025
