Output details
11 - Computer Science and Informatics
University of Birmingham
Regularized Negative Correlation Learning for Neural Network Ensembles
<24>Regularisation was introduced explicitly into negative correlation learning for the first time, whose advantages over existing algorithms were shown through both theoretical derivations and computational studies. It was the 9th most accessed paper in IEEE Transactions on Neural Networks in the month it was published. The work has been followed up by many other researchers and applied to load forecasting in smart grid, industrial process identification, analog curcuit design, and physiological and biomechanical signal classification. The research also led to invited keynote speeches at INES'10 in Spain, ACNNAI'10 in Belarus, ISA'10 in Germany and ICAIS'11 in Austria.