Adaptive Online Sequential ELM for Concept Drift Tackling

Table 4

Average testing accuracy and Cohen’s kappa in % for MNIST VD experiment (other ELM parameters are same: ROS, , ) with 10x trials.

(a) Benchmark result, nonadaptive OS-ELM and offline ELM

Performance	ELM method	()	()	()
Testing accuracy	OS-ELM	95.32 ± 0.12	94.64 ± 0.15	96.86 ± 0.13
	Offline ELM	95.33 ± 0.13	94.66 ± 0.15	96.85 ± 0.06
Cohen’s kappa	OS-ELM	94.80 (0.24)	94.04 (0.25)	96.51 (0.19)
	Offline ELM	94.81 (0.23)	94.06 (0.25)	96.50 (0.19)

(b) VD experiment, AOS-ELM ()

Drift	Testing accuracy	Cohen’s kappa
MNIST   MNIST	96.15 ± 0.08	95.72 (0.21)