Graduate Thesis Or Dissertation
 

A modification of Veto logic for a committee of threshold logic units and the use of 2-class classifiers for function estimation

Público Deposited

Conteúdo disponível para baixar

Baixar PDF
https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/g158bm777

Descriptions

Attribute NameValues
Creator
Abstract
  • The well-known local adjustment algorithm for training a threshold logic unit, TLU, is extended to a local adjustment algorithm for training a network of TLUs Computer simulations show that the extension is unsatisfactory. A new logic for a committee of TLUs, called modified veto logic, and a local adjustment algorithm for training a modified veto committee are described. Unlike a majority committee, a modified veto committee may have members added during training, and the modified veto committee is free to attain a size needed to solve a problem. Computer simulations show that a modified veto committee can solve difficult pattern recognition problems and, in the instances tested, does so more successfully than a majority committee. A technique for using a number of 2-class classifiers to perform function estimation is described. The 2-class classifiers are trained on a set of ordered pairs belonging to the function being estimated; no information about the form of the function is needed; the function can have any number of independent variables; and the accuracy of the estimate increases with the number of 2-class classifiers used. Computer simulations on artificially generated data and on "real life" data show that this technique provides accurate estimates of functions. It is shown that replacing non-binary variables by binary variables can greatly increase the recognition rate of a TLU.
Resource Type
Date Available
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Academic Affiliation
Non-Academic Affiliation
Subject
Declaração de direitos
Publisher
Peer Reviewed
Language
Digitization Specifications
  • File scanned at 300 ppi (Monochrome) using Capture Perfect 3.0.82 on a Canon DR-9080C in PDF format. CVista PdfCompressor 4.0 was used for pdf compression and textual OCR.
Replaces

Relações

Parents:

This work has no parents.

Em Collection:

Itens

Miniatura Título Data de carga Acesso Ações
file details OsborneMartin1975.pdf 2017-08-17 Público Baixar