INTRODUCTION TO THE TLS EXIN NEURON - Neural-based Orthogonal Data Fitting: The EXIN Neural Networks - page 129

Figure 4.4 κ = 0 . 4; averaged over 80 independent trials for w 3 .

Figure 4.5 κ = 0 . 4; averaged over 80 independent trials for w 4 .

− 8and − 26 dB). Figure 4.10 shows the index parameter ρ for κ = 0 . 2. The
difference between the two learnings is smaller because, with less noise, the
TLS GAO linearization is more acceptable. In the last simulations, the additive
input and output noises are zero-mean white Gaussian noise with variance σ

2 .

2

Figure 4.11 shows the case for σ = 0 . 1: TLS EXIN shows an improvement of