1887

n Suid-Afrikaanse Tydskrif vir Natuurwetenskap en Tegnologie - Neurale netwerke as moontlike woordafkappingstegniek vir Afrikaans : research and review article

Volume 22, Issue 1
  • ISSN : 0254-3486
  • E-ISSN: 2222-4173

Abstract

In Afrikaans word saamgestelde woorde aanmekaar geskryf en nuwe woorde word voortdurend gevorm. Aangesien daar dus nie 'n statiese verwysingsbron bestaan nie, is die proses van woordafkapping tydens teksprosessering 'n probleem - veral waar smal kolomme gebruik word, soos in tydskrifte en koerante. 'n Neurale netwerk (vorentoevoer-terugpropagering) is vir die afkappingsprobleem ontwikkel en met sowat 5 000 Afrikaanse woorde met korrekte lettergreepverdeling afgerig. Die neurale netwerk het gemiddeld 97, 56% van moontlike posisies in 5 000 willekeurig gekose woorde korrek as óf geldige óf ongeldige afkappingspunte geklassifiseer. Tydens 'n toets met woorde uit 'n Afrikaanse tydskrif het die neurale netwerk 98, 75% van woordposisies korrek geklassifiseer. Hieruit is die gevolgtrekking gemaak dat neurale netwerke wel suksesvol as afkappingstegniek vir Afrikaans gebruik kan word.

<b>Neural networks as possible hyphenation technique for Afrikaans</b> <br>In Afrikaans compound words are written as one word. New words are therefore created by simply joining words. Word hyphenation during typesetting by computer is often a problem, because the source of reference changes all the time. A neural network (feedforward backpropagation) was trained with about 5 000 Afrikaans words with correct syllabification. The neural network classified 97, 56% of possible points in 5 000 randomly chosen words correctly as either valid or invalid hyphenation points. In a test with 510 words from an Afrikaans magazine the neural network classified 98, 75% of possible positions correctly. We came to the conclusion that neural networks can be used successfully as hyphenation technique for Afrikaans.

Loading full text...

Full text loading...

Loading

Article metrics loading...

/content/aknat/22/1/EJC20327
2003-03-01
2019-12-09

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error