Computational aids for Zulu natural language processing

Pretorius, Laurette; Bosch, Sonja E.

doi:10.2989/16073610309486348

Cited by 9 publications

(7 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This suggests that more careful attention to sentence structure is necessary when selecting content for lessons whether in class or in an app. 6 Indeed, this has also provided challenges to attempts to automate morphological analysis of Zulu (Pretorius & Bosch, 2003).…”

Section: Considering Some Challenges With Learning Zulumentioning

confidence: 99%

The CALL of Zulu: reflections on the development of a computer-assisted language learning package

Christiansen

Els

2019

Computer Assisted Language Learning

View full text Add to dashboard Cite

Few people who did not grow up speaking Zulu have learned the language later. There are limited resources for second language Zulu learning, whether textbooks, readers, or computerised resources. We set out to develop software for this purpose, to support learners' independent learning. Drawing on research on language learning, we used a number of principles that then informed the design of the programmes. In this paper, we reflect on the applicability of the principles and the difficulties in structuring an application for an agglutinative language.

show abstract

Section: Considering Some Challenges With Learning Zulumentioning

confidence: 99%

The CALL of Zulu: reflections on the development of a computer-assisted language learning package

Christiansen

Els

2019

Computer Assisted Language Learning

View full text Add to dashboard Cite

show abstract

“…The xfst script is also compiled into a finitestate network. These networks are finally combined by means of the operation of composition into a so-called Lexical Transducer that constitutes the morphological analyser and contains all the morphological information of Zulu, including derivation, inflection, alternation and compounding (Pretorius & Bosch, 2003b).…”

Section: Containing Overgeneration General Computational Approachmentioning

confidence: 99%

Containing overgeneration in Zulu computational morphology

Pretorius

Bosch

2008

Southern African Linguistics and Applied Language Studies

Self Cite

View full text Add to dashboard Cite

The development of a large-coverage, computational morphological analyser for Zulu requires the modelling not only of the regular phenomena often associated with word formation, but also the idiosyncratic behaviour that may occur in Zulu morphology. This paper discusses the application of an existing rule-based, finite-state morphological analyser prototype ZulMorph in semi-automating the mining of available Zulu language corpora for idiosyncratic behaviour. The semi-automated procedure makes provision for bootstrapping the morphological analyser to include newly extracted information from corpora. Of particular interest is also the central role that the machine-readable lexicon plays. The procedure is applied to a Zulu development corpus of 30 000 types and the results are given and discussed.

show abstract

“…This baseline method already achieves more than 97% accuracy on known words (Table 1), but does not handle unknown words very well with a score of only 18.59%. 2 MXPOST is available from ØÔ »» ØÔº ×ºÙÔ ÒÒº Ù»ÔÙ » Û Ø» ÑÜ» ÑÜºØ Öº Þ 3 MBT is available from ØØÔ »» Ð ºÙÚØºÒÐ»×Ó ØÛ Ö º ØÑÐ 4 SVMTool is available from ØØÔ »»ÛÛÛºÐ× ºÙÔ º ×» ÒÐÔ»ËÎÅÌÓÓÐ» Table 1 indicates that the TnT tagger by far exhibits the most efficient processing times of all data-driven taggers 5 . Despite an exhaustive optimization phase on the validation set (which still revealed the default settings to perform the best), the performance of the TnT tagger trails in direct comparison to the other taggers.…”

Section: Experiments: Individual Tagger Performancementioning

confidence: 99%

“…Compared to the baseline tagger, MXPOST achieves an error reduction rate of 72% (54% on known words, 92% on unknown words). 5 Approximate CPU time was measured on a dual 64bit AMD Opteron 2.44GHz system with 6GB RAM.…”

Section: Experiments: Individual Tagger Performancementioning

confidence: 99%

“…In this regard, the EAGLES tagset was adjusted for Setswana [1], a different tagset and suggestions to venture into Transformation-Based Tagging were presented for isiXhosa [2], yet another tagset and a combination of rule-based symbolic tagging and statistical tagging were offered as a corpus-processing tool for Sesotho sa Leboa [3,4], and a prototype finite-state morphological analyzer was developed for isiZulu [5,4].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Data-Driven Part-of-Speech Tagging of Kiswahili

Pauw

Schryver

Wagacha

2006

Text, Speech and Dialogue

View full text Add to dashboard Cite

Abstract. In this paper we present experiments with data-driven part-of-speech taggers trained and evaluated on the annotated Helsinki Corpus of Swahili. Using four of the current state-of-the-art data-driven taggers, TnT, MBT, SVMTool and MXPOST, we observe the latter as being the most accurate tagger for the Kiswahili dataset.We further improve on the performance of the individual taggers by combining them into a committee of taggers. We observe that the more naive combination methods, like the novel plural voting approach, outperform more elaborate schemes like cascaded classifiers and weighted voting. This paper is the first publication to present experiments on data-driven part-of-speech tagging for Kiswahili and Bantu languages in general.

show abstract

Computational aids for Zulu natural language processing

Abstract: The development of finite-state morphological analyser prototypes for a variety of languages belonging to the Bantu language family, based on underlying machine-readable lexicons that conform to common lexical specifications and international standards.

Cited by 9 publications

References 0 publications

The CALL of Zulu: reflections on the development of a computer-assisted language learning package

The CALL of Zulu: reflections on the development of a computer-assisted language learning package

Containing overgeneration in Zulu computational morphology

Data-Driven Part-of-Speech Tagging of Kiswahili

Contact Info

Product

Resources

About