Journal of Approximation Theory 1992

4 papers
In 1992, the Journal of Approximation Theory featured research connecting classical approximation methods with emerging computational fields. A notable focus was the application of ridge functions to neural networks with a single hidden layer, a topic explored by authors such as Charles K. Chui and Man-Chung Yeung. This work highlights the intersection of functional analysis and machine learning architectures during this period. Additional research from this year addressed multidimensional Bernstein-Durrmeyer operators and their inverse theorems in L spaces. The journal also published studies on the relationship between Jackson's theorem and circulant preconditioned Toeplitz systems, indicating an interest in the convergence properties of numerical linear algebra techniques. Furthermore, the 1992 edition included investigations into established mathematical problems, such as a conjecture by Z. Ditzian, reflecting a continued emphasis on foundational theoretical questions.

Papers

Title
Approximation by ridge functions and neural networks with one hidden layer
Inverse theorems for multidimensional Bernstein-Durrmeyer operators in L
Jackson's theorem and circulant preconditioned Toeplitz systems
On a conjecture of Z. Ditzian