Effective Learning in Dynamic Environments by Explicit Context Tracking
ECML '93 Proceedings of the European Conference on Machine Learning
An ensemble approach for incremental learning in nonstationary environments
MCS'07 Proceedings of the 7th international conference on Multiple classifier systems
Adaptive ensemble based learning in non-stationary environments with variable concept drift
ICONIP'10 Proceedings of the 17th international conference on Neural information processing: theory and algorithms - Volume Part I
| Hi-index | 0.00 |
Virtual concept drift is a phenomenon frequently arising in applications of machine learning theory. Most commonly, a discardretrain strategy is the only option for dealing with newly generated data coming from previously unknown areas of the input space. This paper proposes a method of constructing classifier ensembles based on a measure of observations' density and homogeneity of their corresponding labels. The strategy allows to incrementally add new data points into the model without the necessity of a full retraining procedure.