We analyse errors and deviations made in Basque to bring contributions to two research fields: in the field of automatic error treatment, our aim is to develop spelling, grammar and style checkers for Basque; in the field of ICALL, we aim to create resources for the study of the language learning process. In both fields, in order to develop useful tools which consider the needs that real users might have, it is necessary to carry out a complete analysis of the errors and deviations made in Basque, taking also into account the sociolinguistic situation of our language community.
For that, in this dissertation we have developed a whole environment composed of those resources which are essential for error analysis: corpora (learner and native speaker corpora), an error editor tool called EtikErro, an error classification and two data-bases (named Errors and Learners) to store information related to the tagged examples. We have also defined the basic criterion for the analysis and processing of errors. As it is not possible to study all the error types at the same time, in this dissertation we have dealt with the analysis and automatic treatment of determiner-errors in Basque. We have written some rules that will be integrated in XUXENg, the grammar-checker we are developing for Basque.
ixa.si.ehu.es/Ixa/Argitalpenak/Tesiak/1260962506/publikoak/TESIA-Larraitz_2009
[…] our colleague Larraitz Uria’s PhD thesis was also nominated for this […]
[…] Larraitz Uria’s PhD Thesis: A framework for the analysis of errors and deviations in Basque texts.… […]