<![CDATA[Depression is a mental health problem frequently experienced by students due to various academic, social, and economic demands. Early identification is necessary so that at-risk students can receive faster treatment. This study aims to develop a predictive model for depression indications in students at Bina Sarana Informatika University using the Decision Tree C4.5 algorithm based on data from the Patient Health Questionnaire-9 (PHQ-9) instrument. The study was conducted by following the Cross Industry Standard Process for Data Mining (CRISP-DM) stages, which include understanding the problem, preparing the dataset, building a classification model, and evaluating the model's performance. The research dataset was obtained from the results of filling out the Patient Health Questionnaire-9 (PHQ-9) questionnaire by Bina Sarana Informatika University students. The instrument consists of nine indicators used to measure depressive symptoms based on the frequency experienced by respondents. Respondents with a total score of at least 10 are classified as indicated as depressed, while respondents with a score below 10 are categorized as not indicated as depressed. The dataset was divided into 80% training data and 20% testing data, resulting in an accuracy of 81.40%. Validation was then performed using 10-Fold Cross Validation. Model evaluation was performed using a confusion matrix, precision, recall, F1-score, and Area Under the Curve (AUC). The results showed that the model achieved an average accuracy of 84.55%, indicating that the C4.5 Decision Tree demonstrated good ability to classify depression in college students. Furthermore, the attribute with the highest Information Gain value was the primary factor in the decision tree formation process, thus enabling it to be used as a basis for developing a data mining-based early detection system.]]>
Copyrights © 2026
