DLLog: An Online Log Parsing Approach for Large-Scale System

Table 6

Accuracy comparison of parsing results.

Dataset	Drain	Spell	Nulog	IPLoM	Logram	Brain	DLLog
HDFS	0.998	1	0.998	1	0.981	0.997	1
Hadoop	0.948	0.778	0.954	0.954	0.965	0.915	0.992
Spark	0.920	0.905	0.927	0.920	0.903	0.997	0.995
ZooKeeper	0.967	0.964	0.990	0.955	0.725	0.987	0.979
OpenStack	0.733	0.764	0.989	0.871	0.545	1	0.831
BGL	0.963	0.787	0.954	0.939	0.788	0.978	0.988
HPC	0.887	0.654	0.912	0.824	0.824	0.727	0.996
Thunderbird	0.955	0.844	0.912	0.663	0.761	0.971	0.783
Windows	0.997	0.989	0.990	0.567	0.957	0.718	0.992
Linux	0.690	0.605	0.468	0.672	0.461	0.480	0.672
Mac	0.787	0.757	0.826	0.673	0.673	0.952	0.856
Android	0.911	0.919	0.822	0.712	0.848	0.817	0.905
HealthApp	0.780	0.639	0.876	0.822	0.822	0.875	0.990
Apache	1	1	1	1	1	1	1
OpenSSH	0.788	0.554	0.603	0.802	0.553	0.819	0.751
Proxifier	0.526	0.527	0.515	0.515	0.504	0.526	0.527
Average	0.865	0.792	0.858	0.806	0.783	0.859	0.891

The bold values represent the best comparison result.