Exploring constructive algorithms with stopping criteria to produce accurate and diverse individual neural networks in an ensemble
Md. Monirul Islam, Md Shahjahan, Kazuyuki Murase · 2002
Explores the use of constructive algorithms with stopping criteria to design accurate and diverse individual neural networks (NNs) in an ensemble. Based on constructive algorithms and stopping criteria, a constructive neural network ensemble (CNNE) method to design accurate and diverse individual NNs in an ensemble is proposed. The CNNE is applied to three benchmarked problems in NNs. They are cancer, diabetes and heart disease problems. The experimental results demonstrate the effectiveness of CNNE.