Classification of Psychological Conditions of Informatics Students Using Categorical Naive Bayes within CRISP-DM Framework
DOI:
https://doi.org/10.37012/jtik.v12i2.4151Abstract
This study aims to classify the psychological conditions of Informatics students at University ABC using Categorical Naive Bayes within CRISP-DM Framework. Primary data were collected via a Google Form questionnaire distributed to students, yielding 100 valid respondents with no duplicates. Each respondent answered 10 questions covering academic aspects (task workload, study hours, course difficulty), physical aspects (sleep duration, exercise frequency, health condition), and social aspects (family support, leisure time, peer relationships, gender). Class labels consisted of three categories: Low, Moderate, and High. Data were split at an 80:20 ratio for training and testing and evaluated using 5-Fold Stratified Cross-Validation. Cross-validation results showed that Categorical Naive Bayes achieved a mean Accuracy of 55.0% (±11.4%) and Macro F1 of 39.2% (±13.0%), competitive against Decision Tree (38.4%) and Logistic Regression (37.6%). The model achieved the highest mean Macro F1 among the algorithms compared, although a paired t-test found the difference not statistically significant (p > 0.05), and the margin above the Dummy Classifier (24.7%) remained modest. On the final single-split evaluation, the model achieved Accuracy of 50.00% with Macro F1 of 32.44%. The primary limitation was failure to detect the High class (Recall 0%) due to severe class imbalance (only 13% of data labeled High). Data distribution revealed that the majority of students (59%) experienced Moderate stress levels, 43% slept fewer than 5 hours per night, and 79% reported very good family support.
Downloads
Published
Issue
Section
Citation Check
License
Copyright (c) 2026 Daffa Maulana, Annisa Sakinah Hasibuan, Nadiza Lediwara

This work is licensed under a Creative Commons Attribution 4.0 International License.
Jurnal Teknologi Informatika dan Komputer allows readers to read, download, copy, distribute, print, search, or link to the full texts of its articles and allow readers to use them for any other lawful purpose. The journal allows the author(s) to hold the copyright without restrictions. Finally, the journal allows the author(s) to retain publishing rights without restrictions Authors are allowed to archive their submitted article in an open access repository Authors are allowed to archive the final published article in an open access repository with an acknowledgment of its initial publication in this journal.

Jurnal Teknlogi Informatika dan Komputer is licensed under a Creative Commons Attribution 4.0 International License.









