Convergence monitoring in generalised Hebbian learning
Andrew P. Papliński · 2002
We discuss monitoring of the convergence of the generalised Hebbian learning algorithm which performs principal component analysis or the Karhunen-Loeve transform. We start with presentation of the details of the internal structure of the algorithm. Next, we consider a number of inter-related convergence conditions. Finally, we present example of image coding algorithm in which those conditions are verified.