Multioutput Regression Neural Network Training via Gradient Boosting

SEYEDSAMAN EMAMI, Gonzalo Martínez Muñoz

Abstract: This study proposes a sequential training method for multi-output regression neural networks based on Gradient Boosting. The network is built iteratively, with each component learning to correct previous errors, forming an additive model. Experimental results show that this approach outperforms standard neural network training.