Online learning strategies for classification of static data streams
Mónica Millán-Giraldo, J. Salvador Sánchez · 2008
This paper addresses the problem of online learning for static streaming data. The ultimate objective is to compare a number of very simple learning methods, mainly taken from the literature. We include a straightforward time-weighted strategy for forgetting obsolete objects from the reference set. Experiments are conducted on ten real data sets and using five different classifiers in order to identify which online learning model is the most suitable in terms of classifier performance.