Table 1.
Correlations and descriptive statistics among all measures.
Table 2.
Summary of significant relationships from hierarchical regression modeling.
Table 3.
Performance metrics across models and behaviors for aggregate FoMO.
Table 4.
Performance metrics across models and behaviors for individual FoMO items.
Fig 1.
Comparing aggregate and individual FoMO item performance metrics.
Comparison of model accuracy for the aggregate scenario vs. the individual scenario across behavior domains based on values shown in Tables 3 and 4. Results are aggregated by machine learning model and scenario, with solid bars representing accuracy values for the aggregate scenario and the bars with patterns showing the results for the individual scenario.
Fig 2.
Decision tree output for drug use.
Decision tree for drug offense/use classification based on the FoMO aggregate scenario. Starting at the root node, an example is evaluated in a sequential manner down the tree based on the conditions in the decision nodes. A classification is made according to the end node reached (blue denotes a positive prediction and light orange a negative prediction).
Table 5.
Mean feature importance scores across behaviors.