New
HealthMore in Health→
Machine Learning Flags Low Resilience in Health Care Workers With 75% Accuracy
Key Takeaways
- A study in Discover Artificial Intelligence compared logistic regression, random forest and SVM for classifying health care workers by psychosocial resilience level.
- Logistic regression was most accurate at 75.6% with an area under the ROC curve of 0.816, ahead of random forest (72.6%) and SVM (70.8%).
- The models used the How Right Now Mental Health & Coping dataset from NORC at the University of Chicago, covering 2,055 U.S. respondents surveyed between 2021 and 2022.
- A resilience index was built from four psychometric variables — resilience, ability to bounce back, control and confidence — and split into high and low levels at the median.
DE
DT Editorial Team··via medicalxpress.com