Adaptive Online Sequential ELM for Concept Drift Tackling

Table 1

Concept drift scenarios, compared methods, and sequential patterns.

(a) The experiment design scenarios

Data set	Virtual drift	Real drift	Hybrid drift	Compared methods
SEA	—	✓	—	OS-ELM, CEOS-ELM, Kolter [20]
STAGGER	—	✓	—	OS-ELM, CEOS-ELM, Kolter [20]
MNIST	✓	✓	✓	OS-ELM, Offline ELM, ELM ensemble
MNIST + USPS	—	✓	✓	OS-ELM, offline ELM, ELM ensemble

(b) Concept drift sequential patterns

Data set	Sequential patterns scenarios	Cause of shift
SEA	Sudden change	Linear discriminant function
STAGGER	Sudden change	Logical discriminant rule
MNIST	Sudden change and recurring context	Additional attributes or classes
USPS	Recurring context	Additional attributes or classes