Title :
On the Additive Properties of the Fat-Shattering Dimension
Author :
Asor, Ohad ; Duan, Hubert Haoyang ; Kontorovich, Aryeh
Author_Institution :
Adv. Comput. R&D, Rehovot, Israel
Abstract :
The properties of the VC-dimension under various compositions are well-understood, but this is much less the case for classes of continuous functions. In this brief, we show that a commonly used scale-sensitive dimension, Vy, is much less well-behaved under Minkowski summation than its VC cousin, while the fat-shattering dimension retains some compositional similarity to the VC-dimension. As an application, we analyze the fat-shattering dimension of trigonometric functions and series.
Keywords :
learning (artificial intelligence); series (mathematics); Minkowski summation; VC-dimension; additive property; continuous functions; fat-shattering dimension; scale-sensitive dimension; trigonometric functions; trigonometric series; Additives; Boosting; Complexity theory; Convergence; Neural networks; Upper bound; Combinatorial dimension; Minkowski addition; fat-shattering; scale-sensitive; scale-sensitive.;
Journal_Title :
Neural Networks and Learning Systems, IEEE Transactions on
DOI :
10.1109/TNNLS.2014.2327065