Selective negative correlation learning algorithm for incremental learning

Negative correlation learning (NCL) is a successful scheme for constructing neural network ensembles. In batch learning mode, NCL outperforms many other ensemble learning approaches. Recently, NCL is also shown to be a potentially powerful approach to incremental learning, while the advantage of NCL has not yet been fully exploited. In this paper, we propose a selective NCL approach for incremental learning. In the proposed approach, the previously trained ensemble is cloned when a new data set presents and the cloned ensemble is trained on the new data set. Then, the new ensemble is combined with the previous ensemble and a selection process is applied to prune the whole ensemble to a fixedsize. Simulation results on several benchmark datasets show that the proposed algorithm outperforms two recent incremental learning algorithms based on NCL.