Title :
Experiments with Boosted Decision Tree Classifiers
Author_Institution :
Dept. of Syst. & Comput. Networks, Wroclaw Univ. of Technol., Wroclaw
Abstract :
Boosting is the most popular method of improving quality and stabilizing weak classifiers. It bases on the voting by the group of classifiers, where each of them is generated on the basis of modified original learning set. The modification of AdaBoost.M1 and experimental results of boosted C4.5 (decision tree induction) algorithm are presented. All experimental researches are made on well known benchmark databases.
Keywords :
decision trees; learning (artificial intelligence); pattern classification; AdaBoost.M1; boosted decision tree classifiers; decision tree induction algorithm; Application software; Boosting; Classification tree analysis; Computer networks; Decision making; Decision trees; Induction generators; Intelligent networks; Intelligent systems; Voting; AdaBoost; boosting; decision tree;
Conference_Titel :
Intelligent Systems Design and Applications, 2008. ISDA '08. Eighth International Conference on
Conference_Location :
Kaohsiung
Print_ISBN :
978-0-7695-3382-7
DOI :
10.1109/ISDA.2008.215
