Recursive decompounding in Afrikaans

Authors:
Tilla Fick;Chris Swanepoel
Affiliations:
Department of Decision Sciences, University of South Africa, Pretoria, South Africa;Department of Decision Sciences, University of South Africa, Pretoria, South Africa
Venue:
TSD'11 Proceedings of the 14th international conference on Text, speech and dialogue
Year:
2011

Citing 8
Cited 0

Shallow Morphological Analysis in Monolingual Information Retrieval for Dutch, German, and Italian

CLEF '01 Revised Papers from the Second Workshop of the Cross-Language Evaluation Forum on Evaluation of Cross-Language Information Retrieval Systems
Empirical methods for compound splitting

EACL '03 Proceedings of the tenth conference on European chapter of the Association for Computational Linguistics - Volume 1
Introduction to Information Retrieval

Introduction to Information Retrieval
Decompounding query keywords from compounding languages

HLT-Short '08 Proceedings of the 46th Annual Meeting of the Association for Computational Linguistics on Human Language Technologies: Short Papers
Towards better machine translation quality for the German--English language pairs

StatMT '08 Proceedings of the Third Workshop on Statistical Machine Translation
German decompounding in a difficult corpus

CICLing'08 Proceedings of the 9th international conference on Computational linguistics and intelligent text processing
How to avoid burning ducks: combining linguistic analysis and corpus statistics for German compound processing

WMT '10 Proceedings of the Joint Fifth Workshop on Statistical Machine Translation and MetricsMATR
Statistical machine translation of german compound words

FinTAL'06 Proceedings of the 5th international conference on Advances in Natural Language Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

An algorithm has been developed to decompose compound words in Afrikaans. This data driven technique recursively uses an extensive list of Afrikaans words in the decompounding process. String fitting from the beginning and end of words forms the basis of the process, while sublists containing short words that may occur only at the beginning or end of words, and lists of prefixes and suffixes are utilised. Applying the algorithm to the original lexicon of 182 433 words resulted in accuracy of 90,2%, precision of 99,9% and recall of 83,6%.