Situation Element Extraction Based on Fuzzy Rough Set and Combination Classifier

Table 3

Comparison of training and recognition of classifiers.

Classifier	Total recognition rate	Normal recall	Probe recall	DOS recall	U2R recall	R2L recall	Train time (s)
Linear SVC	0.7574	0.6575	0.8469	0.9637	0.8889	0.4444	13.12
SVC	0.7468	0.6541	0.7965	0.9544	0.8333	0.3846	105.23
Complement NB	0.6288	0.6372	0.3795	0.8400	0.0000	0.0000	0.06
MLP	0.7505	0.6534	0.7916	0.9648	0.7222	0.9404	191.03
K-Neighbors	0.7548	0.6602	0.8108	0.9573	0.8462	0.9322	43.62
Logistic regression	0.7588	0.6528	0.8914	0.9690	0.7500	0.5000	54.68
SGD (log)	0.7423	0.6384	0.8776	0.9665	0.0000	0.4444	2.71
SGD (modified_huber)	0.7549	0.6552	0.8466	0.9628	0.6667	0.2222	3.01
Decision tree	0.7724	0.6874	0.8569	0.9209	0.3125	0.7736	2.53
Extra tree	0.7309	0.6540	0.7639	0.9471	0.5385	0.4471	0.30
Gradient boosting	0.7241	0.6357	0.7112	0.9582	0.9167	0.1321	290.57
AdaBoost	0.7241	0.6781	0.5629	0.8576	0.2500	0.8202	15.54
Random forest	0.7527	0.6478	0.8783	0.9643	1.0000	0.9623	23.73
Bagging	0.7429	0.6560	0.7729	0.9216	0.4000	0.8304	18.11
Extra trees	0.7425	0.6429	0.8372	0.9587	0.0000	0.8333	23.65