Purpose: This study aims to implement the Decision Tree C4.5 algorithm to classify bullying and sexual harassment cases in senior high schools and develop a web-based decision support system for consistent, evidence-based identification and intervention.
Methodology: A quantitative experimental approach was applied. Data were collected through anonymous student surveys and interviews with Guidance and Counselling (BK) teachers, resulting in 120 cases (93 bullying and 27 sexual harassment). The C4.5 algorithm was implemented using RapidMiner, while the web system was developed using Waterfall System Development Life Cycle (SDLC) with Personal Home Page (PHP) Laravel, MySQL, HTML/CSS, and tested using black box testing.
Results: The model produced a total entropy of 2.44989, with “Incident Type” as the root node (Information Gain = 1.811). “Incident Frequency” became the second-level node. The system successfully classified cases and provided recommendations with 100% success in all nine black box tests covering authentication, classification, reporting, and data management modules.
Conclusions: The C4.5 algorithm effectively classifies bullying and sexual harassment cases, while the web-based system enhances consistency and reduces subjectivity in school decisionmaking.
Limitations: The dataset is limited to 120 cases at the senior high school level, without precision, recall, or F1-score analysis and no longitudinal data.
Contributions: This study provides an operational decision support system using C4.5 for structured classification of schoolbased bullying and sexual harassment cases.