Data Mining and Machine Learning Applied to Mental Health Prediction in Information Technology Workers

Antunez K.; Cortez I.; Gomez M.; Gonzales A.; Santiago N.; Calderón D.; Cárdenas F. · Computacion y Sistemas (2025) · DOI: 10.13053/cys-29-3-4520

Resumen

Currently, mental health is an increasingly frequent problem. Mental disorders such as anxiety and depression tend to contribute to work-related behavioral changes, such as reduced activity levels and poor performance. The objective of this research is to measure attitudes towards mental health in the IT sector. The ANOVA F-test and Chi-Square filter were used as feature selection techniques. For modeling, the following algorithms were applied: K-Nearest Neighbor (KNN), Decision Tree, Support Vector Machine (SVM), Naive Bayes, and Neural Network. The study used a dataset obtained from the OSMH/OSMI Mental Health in Tech Survey, which included around 1400 responses from a survey conducted in 2016. The Neural Network algorithm achieved the best results on the analyzed dataset. The model metrics obtained were an F1-Score of 85.92%, an Area Under the ROC Curve (AUC) of 0.903, a lower value of 23 false negatives, and a higher value of 119 true positives in the confusion matrix.

Instituciones: Universidad Nacional Mayor de San Marcos