Fig. 4

Six ML algorithms for classification with 39 DEGs. A LASSO for eight candidate genes and the error bars mean the fluctuation range of Binomial Deviance; B SVM for 13 candidate genes. C RF discriminated between the control and AMI groups. And the red, black, and green lines represent the Con, out-of-bag (OOB), and AMI groups respectively. D DT discriminated between the control and AMI groups. E A sixfold GBM submodel was constructed. The heat map illustrates the importance of genes in each respective submodel. The intensity of the color corresponds to the significance of the gene in the particular submodel. F NN discriminated between the control and AMI groups. All 39 DEGs were involved in modelling in NN, and there are ten because of space limitations. If an edge is colored red, it indicates a positive correlation, meaning that the current feature positively affects the classification result. Conversely, if the edge is gray, it implies a negative correlation. Furthermore, the thickness of the edge signifies the weight's magnitude