An Equivalent Form of S-L Projection Learning
Benyong Liu, Ogawa Hidemitsu · Journal of Electronic Science and Technology · 2003
Learning a mapping between an input space and an output space from training data can be viewed as a problem of function approximation, which means that different criteria result in learning approaches of different abilities. Among them, projection criterion proposed by one of the authors aims directly at optimal generalization. The projection idea leads to three specific learning approaches, projection learning, partial projection learning, and averaged projection learning, and a framework of a family of projection learning called S-L projection learning is established to discuss infinite kinds of learning. S-L projection learning has a dijfevent form from the three methods, and it is not easy to analyze their relationship. This paper focuses on an equivalent form of S-L projection learning, which shows that it is a transformed version of partial projection learning.