Comparing Various Tree-Based Model Approaches to Spot Self-Sabotage Behavior
JOIV : International Journal on Informatics Visualization · 31.07.2026 · 10 (4) · 1923
DOI: 10.62527/joiv.10.4.4467 ↗Kurz gefasst
Fachartikel in der Zeitschrift JOIV : International Journal on Informatics Visualization, online erschienen am 31.07.2026. Verfasst von Mochammad Anshori und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).
Abstract (Originalsprache)
This study presents a novel, interpretable machine-learning framework to detect self-sabotaging behavior in university students by systematically comparing six tree-based classifiers on primary questionnaire data. Using responses from 50 students and 11 features (psychological constructs such as imposter syndrome, fear of failure, and self-esteem, plus age and gender), we evaluate Decision Tree, Random Forest, AdaBoost, CatBoost, Gradient Boosting, and XGBoost with min–max normalization, label encoding, and random oversampling to address class imbalance, and 10-fold cross-validation for robust estimation. Unlike prior work that typically applies a single algorithm or focuses on other mental health outcomes, our contribution is (1) a head-to-head comparison of multiple interpretable tree-based methods specifically for self-sabotage detection, (2) an explicit assessment of how class-balancing affects each algorithm, and (3) practical recommendations for deploying interpretable models in small, mixed-type psychological datasets. Key results show that CatBoost achieved the best post-oversampling performance (accuracy = 87.67%, precision = 0.80, recall = 0.799, F1-score = 0.80), while Gradient Boosting scored highest on the original imbalanced data (accuracy = 77.67%). These findings highlight the importance of preprocessing, especially oversampling, and the suitability of CatBoost for mixed categorical/numerical psychological data. We conclude that interpretable tree-based models can provide reliable, scalable screening tools for the early identification of self-sabotaging tendencies; future research and other resampling techniques, such as SMOTE or feature selection to assess feature relevance, are recommended.
Zitieren
APA
Anshori, M., Haris, M. S., & Kusuma, W. T. (2026). Comparing Various Tree-Based Model Approaches to Spot Self-Sabotage Behavior. JOIV : International Journal on Informatics Visualization, 10(4), 1923. https://doi.org/10.62527/joiv.10.4.4467
Vancouver
Anshori M, Haris MS, Kusuma WT. Comparing Various Tree-Based Model Approaches to Spot Self-Sabotage Behavior. JOIV : International Journal on Informatics Visualization. 2026;10(4):1923. doi:10.62527/joiv.10.4.4467
BibTeX
@article{Anshori2026,
author = {Anshori, Mochammad and Haris, M Syauqi and Kusuma, Wahyu Teja},
title = {Comparing Various Tree-Based Model Approaches to Spot Self-Sabotage Behavior},
journal = {JOIV : International Journal on Informatics Visualization},
year = {2026},
volume = {10},
number = {4},
pages = {1923},
doi = {10.62527/joiv.10.4.4467}
}
EUVIMED