The Complete SPRLIB & ANNLIB

The Complete SPRLIB & ANNLIB

Second edition, Version 3.1

CIP-DATA KONINKLIJKE BIBLIOTHEEK, DEN HAAG

Hoekstra, Aarnoud
Kraaijveld, Martin Alfred
De Ridder, Dick
Schmidt, Wouter Frederik

The Complete SPRLIB & ANNLIB
Hoekstra, Aarnoud
Kraaijveld, Martin Alfred
De Ridder, Dick
Schmidt, Wouter Frederik

ISBN 90-75691-03-3
NUGI 841
Subject headings: Statistical Pattern Recognition / Artificial Neural Networks

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means - electronic, mechanical, by photocopying, recording, or otherwise - without the prior written permission of the authors.

Copyright ©1996 by A. Hoekstra, M.A. Kraaijveld, D. de Ridder, W.F. Schmidt

This book was typeset using the LaTeXtypesetting software and printed in The Netherlands at the Delft University of Technology.

• The SPRANNLIB user guide
  • Introduction to SPRANNLIB
    • Introduction
    • System requirements
    • SPRANNLIB implementation
    • Overview of the manual
  • Statistical pattern recognition library
    • Important data structures
      • Vectors and matrices
      • Data sets
    • Important support functions in SPRLIB
      • Additional matrix and vector functions
      • Supporting data set functions
    • Pattern recognition algorithms in SPRLIB
  • Artificial neural network library
    • Considerations on Network Representation
    • Implementation
      • The top level network data structure
      • The unit structure
      • The link data structure
      • The weight structure
      • The unit value structure
    • Functions and Algorithms
      • Considerations on functions and algorithms
      • Supporting network functions
    • Learning algorithms
      • Feed forward networks
      • Kohonen networks
  • Tools
    • Interfacing to other software packages
    • Backpropagation program
    • Generic application set-up
  • An example application
• References
  • Data I/O formats
    • Dataset I/O format
    • Network I/O format
    • Recipe I/O format
      • Recipe-files
      • Input layer
      • Feature maps and subsampling maps
      • Hidden layer
      • Output layer
  • Supported computer platforms
    • Sun workstations
      • Compiling the library
    • Silicon Graphics
      • Compiling the library
    • HP workstations
      • Compiling the library
    • IBM PC and compatibles
      • MS-DOS and Turbo C(++)
      • MS-DOS and DJGPP (GNU C)
      • Windows 95/NT and commercial C-compilers
      • Linux
• The SPRANNLIB reference manual
• About this document ...