Recurring Concepts and Meta-learning

Varování

Publikace nespadá pod Ekonomicko-správní fakultu, ale pod Fakultu informatiky. Oficiální stránka publikace je na webu muni.cz.
Autoři

GAMA Joao KOSINA Petr

Rok publikování 2010
Druh Článek ve sborníku
Fakulta / Pracoviště MU

Fakulta informatiky

Citace
www http://www.ifi.uzh.ch/ddis/fileadmin/pdf/kietz/Proc-3rd-PlanLearn-ECAI-ws09.pdf
Klíčová slova Data Streams; Concept Drift; Meta-learners; Recurrent Concepts
Popis This work addresses data stream mining from dynamic environments where the distribution underlying the observations may change over time. In these contexts, learning algorithms must be equipped with change detection mechanisms. Several methods have been proposed able to detect and react to concept drift. When a drift is signaled, most of the approaches use a forgetting mechanism, by releasing the current model, and start learning a new decision model. Nevertheless, it is not rare for the concepts from history to reappear, for example seasonal changes. In this work we present a method that memorizes learnt decision models whenever a concept drift is signaled. The system uses meta-learning techniques that characterize the domain of applicability of previous learnt models. The meta-learner can detect re-occurrence of contexts and take pro-active actions by activating previous learnt models. The main benefit of this approach is that the proposed meta-learner is capable of selecting similar historical concept, if there is one, without the knowledge of true classes of examples.
Související projekty:

Používáte starou verzi internetového prohlížeče. Doporučujeme aktualizovat Váš prohlížeč na nejnovější verzi.