Machine learning regression models for internal shame

Nataša Kovač, Kruna Ratković, Hojjatollah Farahani, Peter Watson,
Machine learning regression models for internal shame,
Acta Psychologica,
Volume 260,
2025,
105721,
ISSN 0001-6918,
https://doi.org/10.1016/j.actpsy.2025.105721.
(https://www.sciencedirect.com/science/article/pii/S0001691825010340)
Abstract: This study aims to predict Internal Shame (IS) based on childhood trauma, social emotional competence, cognitive flexibility, distress tolerance and alexithymia in an Iranian sample. The regression results suggested that distress tolerance was the most significant predictor, whereas cognitive flexibility had the least impact. We initially tested nine machine learning regression techniques (Multi-Layer Perceptron, AdaBoost, Support Vector Regression, Artificial Neural Network, Decision Tree, Random Forest, Gradient Boosting, Stochastic Gradient Boosting, and Extreme Gradient Boosting). Based on performance evaluation, we retained five models (Decision Tree, Random Forest, Gradient Boosting, Stochastic Gradient Boosting, and XGBoost) for detailed analysis of IS. The findings indicate that the XGBoost regression model was superior in performance compared to the other applied methods.
Keywords: internal shame; regression; machine learning