A Classification and Novel Class Detection Algorithm for Concept Drift Data Stream Based on the Cohesiveness and Separation Index of Mahalanobis Distance

Table 5

Experimental result data in Covertype dataset.

C	C&NCBM accuracy (%)	MineClass accuracy (%)	KNN accuracy (%)	C&NCBM evaluation time (s)	MineClass evaluation time (s)	KNN evaluation time (s)
1	88.9832	87.8438	87.0177	21.5172	18.2156	15.7438
2	91.1561	89.8646	89.4737	39.0094	32.5750	30.7797
3	91.5212	89.8546	89.5844	57.2563	49.2234	46.3016
4	92.0270	90.2581	89.9857	75.3641	65.4875	64.8359
5	91.4612	89.4413	88.9274	96.3563	85.3375	84.2594
6	91.4837	89.2808	88.8317	118.4125	104.6641	102.9203
7	91.4707	89.1186	88.5925	143.8547	128.4422	121.6672
8	91.7367	89.4059	88.8483	165.1609	148.0813	139.3656
9	91.8901	89.3561	88.9612	185.7063	166.9156	156.5406
10	92.2249	89.8698	89.4625	204.0609	183.3109	173.3313