摘要
选择性集成学习是为解决同一个问题而训练多个基分类器,并依据某种规则选取部分基分类器的结果进行整合的学习算法。通过选择性集成可以获得比单个学习器和全部集成学习更好的学习效果,可以显著地提高学习系统的泛化性能。提出了一种多层次选择性集成学习算法Ada_ens。试验结果表明,Ada_ens具有更好的学习效果和泛化性能。
Selective ensemble learning is a learning algorithm,which trains a number of base classifiers and selects part of their outcomes according to a certain rule to assemble.With the selective ensemble,this algorithm would achieve better learning effect than the single classifier and the entire ensemble learning.The algorithm can also remarkably improve the generalisation property of the learning system.In this paper,we propose a multi-level selective ensemble learning algorithm,Ada_ens.Test results show that Ada_ens has better learning results and higher ability to generalize.
出处
《计算机应用与软件》
CSCD
2011年第1期16-18,共3页
Computer Applications and Software
基金
国家自然科学基金项目(60473142)
关键词
机器学习
集成学习
选择性集成
Machine learning Ensemble learning Selective ensemble
