In progress 2026
Burnout prediction with synthetic data
Classification model to estimate burnout risk from Likert-type questionnaires, focusing on calibration, interpretability (SHAP) and group-wise bias analysis.
Data science and machine learning projects applied to psychology.
In progress Classification model to estimate burnout risk from Likert-type questionnaires, focusing on calibration, interpretability (SHAP) and group-wise bias analysis.
Completed Factor analysis and IRT models to validate a short well-being scale, with reproducible R and Python code.
Completed Text classification and topic modeling on a public corpus, with a discussion of the ethical limits of inferring mental states from language.