Webshell Detection Based on Executable Data Characteristics of PHP Code

Table 6

The evaluation index results.

Model name	Classification algorithm	Evaluation index
		Accuracy	Precision	Recall	F1 score
Model 1	Random forest	0.9966	0.9897	0.9954	0.9926
	Support vector machine	0.9942	0.9841	0.9908	0.9874
	Multilayer perceptron	0.9953	0.9863	0.9931	0.9897
Model 2	Random forest	0.9926	0.9896	0.9782	0.9839
	Support vector machine	0.9792	0.9853	0.9233	0.9533
	Multilayer perceptron	0.9871	0.9725	0.9714	0.9719
Model 3	Random forest	0.9942	0.9819	0.9931	0.9875
	Support vector machine	0.9884	0.9621	0.9885	0.9751
	Multilayer perceptron	0.9918	0.9806	0.9840	0.9823

Comparing the same algorithm in different models, the maximum value of the indicator.