作者: Akira Hirose , None
DOI: 10.23919/APSIPA.2018.8659590
关键词:
摘要: This paper discusses the advantages in complex-valued and quaternion neural networks (CVNNs QNNs) by presenting important concepts as well applications electronic engineering. Essentially, CVNNs QNNs have excellent generalization ability learning of two- three-dimensional rotation scaling, respectively. Its origin is sparsity themselves complex numbers. The useful particular adaptive processing waves consisting amplitude (amplitude phase) and/or polarization. We consider an example communications where we deal with channel prediction to discuss superiority its basics.