An Adaptive Distributed Ensemble Approach to Mine Concept-drifting Data Streams.pdfVIP

Progetto: LOGICA Laboratorio Tecnologico della Logistica in Calabria Laboratorio di Cosenza Modulo Realizzativo MR4-INFO Applicazione di modelli avanzati di ottimizzazione/simulazione ai sistemi logistici avanzati Obiettivo Realizzativo OR4-INFO-ICAR Tecniche per la gestione di processi logistici e l’analisi dei dati Rapporto Tecnico An Adaptive Distributed Ensemble Approach to Mine Concept-drifting Data Streams A cura di Gianluigi Folino, Clara Pizzuti, Giandomenico Spezzano, 20 Aprile 2007 Codice: RT.OR4.INFO.ICAR/5 RAPPORTO TECNICO RT.OR4.INFO.ICAR/5 Pagina 1 di 12 INDICE 1. Introduction ______________________________________________________________ 3 2. Gp Boosting ensemble for distributed data streams _______________________________5 2.1 Classification througt Genetic Programming _______________________________ __5 2.2 The StreamGP Method ________________________________________________ ___6 3. Experimental Results _______________________________________________________ __8 4. Conclusions ______________________________________________________________ _11 An Adaptive Distributed Ensemble Approach to Mine Concept-drifting Data Streams Gianluigi Folino, Clara Pizzuti, and Giandomenico Spezzano Institute for High Performance Computing and Networking, CNR-ICAR, Via P. Bucci 41C 87036 Rende (CS), Italy {folino,pizzuti,spezzano}@r.it Abstract. An adaptive boosting ensemble algorithm for classifying ho- mogeneous distributed data streams is presented. The method builds an ensemble of classifiers by using Genetic Programming (GP) to in- ductively generate decision trees, each trained on different parts of the distributed training set. The approach adopts a co-evolutionary architec- ture to support a cooperative model of GP. A change detection strategy, based on self-similarity of the ensemble behavior, and measured by its fractal dimension, permi