Figure 7.

Thirty potential biomarker combinations screened based on operational taxonomic unit (OTU) abundance data. The mean decrease accuracy and the mean decrease Gini coefficient are two important indicators in the random forest model. Mean decrease accuracy indicates the degree of decrease in the accuracy of the random forest prediction; a larger value indicates a greater importance of the variable. Mean decrease Gini coefficient calculates the difference between each variable and the observed value on each node of the classification tree. Qualitative effects, used to compare the importance of variables, indicate that, the greater the value, the greater the importance of the variable.