Algorithm 1 Feature Selection.

Input: Training set with permissions and intent static features;

Output: Optimal subset of features after selection; 1

1:  Step1: Feature importance ranking

2:  The feature importance scores were calculated using chisquare test, analysis of variance Fvalue, and mutual information, respectively.;

3:  Removal of features with scores of Nan and 0;

4:  Obtain the corresponding candidate feature sets separately;

5:  EndStep

6:  Step2: Comparing the average performance of different algorithms

7:  Apply some detection algorithms to the three candidate feature sets;

8:  Calculate the average performance of each algorithm on the three feature sets;

9:  EndStep

10:  Step3: Obtain the optimal subset of features

11:  Compare the average performance and find the best-performing detection algorithm;

12:  Compare the performance of this optimal detection algorithm on three feature subsets and find the best performing feature set;

13:  EndStep

14: Return the optimal subset of features.