An Intrusion Detection Method Based on Decision Tree-Recursive Feature Elimination in Ensemble Learning

Table 4

Features of DoS, Probe, R2L, and U2R on NSL-KDD

DoS	'dst_host_same_srv_rate'; 'dst_host_serror_rate'; 'num_compromised'; 'same_srv_rate'; 'diff_srv_rate'; 'dst_host_count'; 'dst_host_srv_serror_rate'; 'ecr_i'; 'RSTR'; 'wrong_fragment'; 'dst_bytes'; 'src_bytes.'
Probe	'src_bytes'; 'telnet'; 'smtp'; 'private'; 'http'; 'ftp_data'; 'finger'; 'dst_host_rerror_rate'; 'dst_host_same_src_port_rate'; 'dst_host_diff_srv_rate'; 'dst_host_same_srv_rate'; 'rerror_rate'; 'dst_bytes.'
R2L	'duration'; 'imap4'; 'ftp_data'; 'dst_host_srv_diff_host_rate'; 'dst_host_same_src_port_rate'; 'dst_host_same_srv_rate'; 'dst_host_srv_count'; 'dst_host_count'; 'num_access_files'; 'num_failed_logins'; 'hot'; 'dst_bytes'; 'src_bytes.'
U2R	'duration'; 'dst_host_srv_diff_host_rate'; 'dst_host_count'; 'srv_count'; 'num_shells'; 'num_file_creations'; 'root_shell'; 'dst_bytes'; 'dst_host_same_srv_rate'; 'hot'; 'src_bytes.'