paperKB
coga / coga-kb
Help
Sign in

Chunk #26 — 3. Results — 3.1. Random Forests Classification — 3.1.3. Multi-Way Importance

Source
Differentiating Individuals with and without Alcohol Use Disorder Using Resting-State fMRI Functional Connectivity of Reward Network, Neuropsychological Performance, and Impulsivity Measures.
Embedded
yes

Text

The top significant variables can also be shown by a multi-way importance plot based on any of the three RF importance measures. Figure 5 illustrates the important features that contributed to group classification based on Gini decrease, number of trees, and p-value. These features include: (i) 12 FC variables (representing connections across cortical and subcortical regions) (see Table 2 for details), (ii) one neuropsychological variable (i.e., total number of correct trials in forward span), and (iii) two impulsivity scores (motor and non-planning categories).