ASCE will be sunsetting the Civil Engineering Database (CEDB) in the coming months. We encourage you to start using the ASCE Library to perform content searches. In the ASCE Library, you can save your search results as well as receive email content alerts when new content relevant to your area of interest is published. Try it out today at https://ascelibrary.org.

 

Empirical Performance Comparison for Symbolic and Subsymbolic Learning Systems

by Tomasz Arciszewski, George Mason Univ, Fairfax, United States,
Adam Borkowski, George Mason Univ, Fairfax, United States,
Tomasz Dybala, George Mason Univ, Fairfax, United States,
Janusz Racz, George Mason Univ, Fairfax, United States,
Piotr Wojan, George Mason Univ, Fairfax, United States,



Document Type: Proceeding Paper

Part of: Computing in Civil Engineering

Abstract:

The paper gives a description of a comparative study of two learning systems based on symbolic and subsymbolic learning. The two systems' performance was analyzed using the overall empirical error rate. The performance was determined for individual stages in a multi-stage knowledge acquisition process, and learning curves were produced. The comparison was conducted using the same collection of 332 examples of optimal (minimum weight) designs of wind bracings in steel skeleton structures of tall buildings.

Subject Headings: Steel structures | Comparative studies | Structural analysis | Minimum weight design | High-rise buildings | Bracing | Wind engineering

Services: Buy this book/Buy this article

 

Return to search