Computational Technology Resources - CCP

H. Lee* and P. Hajela+

*GE Corporate Research and Development, Schenectady, New York, United States of America
+Rensselaer Polytechnic Institute, Troy, New York, United States of America

doi:10.4203/ccp.16.2.2

purchase the full-text of this paper

H. Lee, P. Hajela, "Vector Clustering for Neural Network based Prediction of Geometrical Characteristics", in B.H.V. Topping, A.I. Khan, (Editors), "Neural Networks & Combinatorial Optimization in Civil & Structural Engineering", Civil-Comp Press, Edinburgh, UK, pp 19-29, 1993. doi:10.4203/ccp.16.2.2

Abstract

This paper explores the development of a function approximation strategy that predicts certain properties based on geometrical characteristics. Such approximations are valuable in situations where the actual response computations are CPU-intensive, and a quick estimation is required. Issues discussed include shape representation using FPF transformations, estimation of mapping nonlinearity using dendrograms, as well as shape based mapping using backpropagation (BP) neural networks. Although BP networks have offered an enhanced mapping capacity, in practice, the formulation of a smoother input/output mapping space is still rather critical because a smoother mapping requires fewer samples to characterize and is easier for a network to learn. The methodology is applied to predict two continuous properties for arbitrary shapes, including non-simple connected ones. The two properties are 1) the ratio of moment of inertia and 2) the radius of gyration. These properties are not all related to the sampling location, orientation and size.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £34 +P&P)

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Proceedings ISSN 1759-3433 CCP: 16 NEURAL NETWORKS & COMBINATORIAL OPTIMIZATION IN CIVIL & STRUCTURAL ENGINEERING Edited by: B.H.V. Topping and A.I. Khan Paper II.2 Vector Clustering for Neural Network based Prediction of Geometrical Characteristics H. Lee* and P. Hajela+ *GE Corporate Research and Development, Schenectady, New York, United States of America +Rensselaer Polytechnic Institute, Troy, New York, United States of America doi:10.4203/ccp.16.2.2 purchase the full-text of this paper Full Bibliographic Reference for this paper H. Lee, P. Hajela, "Vector Clustering for Neural Network based Prediction of Geometrical Characteristics", in B.H.V. Topping, A.I. Khan, (Editors), "Neural Networks & Combinatorial Optimization in Civil & Structural Engineering", Civil-Comp Press, Edinburgh, UK, pp 19-29, 1993. doi:10.4203/ccp.16.2.2 Abstract This paper explores the development of a function approximation strategy that predicts certain properties based on geometrical characteristics. Such approximations are valuable in situations where the actual response computations are CPU-intensive, and a quick estimation is required. Issues discussed include shape representation using FPF transformations, estimation of mapping nonlinearity using dendrograms, as well as shape based mapping using backpropagation (BP) neural networks. Although BP networks have offered an enhanced mapping capacity, in practice, the formulation of a smoother input/output mapping space is still rather critical because a smoother mapping requires fewer samples to characterize and is easier for a network to learn. The methodology is applied to predict two continuous properties for arbitrary shapes, including non-simple connected ones. The two properties are 1) the ratio of moment of inertia and 2) the radius of gyration. These properties are not all related to the sampling location, orientation and size. purchase the full-text of this paper (price £20) go to the previous paper go to the next paper return to the table of contents return to the book description purchase this book (price £34 +P&P)
Back to top	©Civil-Comp Limited 2023 - terms & conditions