Proposed Anxiety Classification Model Using Machine Learning Techniques

Fiestas-Arroyo D.; Ticona W. · Lecture Notes in Networks and Systems (2025) · DOI: 10.1007/978-3-032-00239-6_4

Resumen

Currently, anxiety disorders are one of the most common mental health conditions worldwide, and they have brought serious consequences because they were not treated in time because it was not known that the patient suffered from such a condition. Therefore, the objective of this study is to propose a classification model that uses ML techniques to detect the type of anxiety. The applied methodology consisted of 5 phases: Data-set acquisition, Preprocessing, Feature Selection (LightGBM, RF, Gradient-Boosting, Linear SVM, AdaBoost, Logistic Regression), Machine Learning (LR, RF, XGBoost, NB, SVM, MLP). The best results were obtained for the model using the 13 most important features according to Gradient Boosting, together with the XGBoost classification algorithm, with results after tuning with GridSearch (colsample_bytree: 0. 6; learning_rate: 0.2; max_depth: 5; min_child_weight: 1; n_estimators: 150; subsample: 0.6) of Accuracy, 96.88%, Precision, 96.89%, Recall, 96.88%; and F1-Score, 96.88%. The results succeeded in proposing a robust classification model for the classification of anxiety types. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

Instituciones: Universidad Tecnológica del Perú