Exploring the Influence of Spelling Errors on Lexical Variation Measures

Proceedings of the 27th International Conference on Computational Linguistics, pages 2391–2398 Santa Fe, New Mexico, USA, August 20-26, 2018. 2391 Exploring the Influence of Spelling Errors on Lexical Variation Measures Ryo Nagata†‡, Taisei Sato†, and Hiroya Takamura♭♠ †Konan University ‡RIKEN AIP ♭National Institute of Advanced Industrial Science and Technology ♠Tokyo Institute of Technology nagata-coling2018-k @ hyogo-u.ac.jp. Abstract This paper explores the influence of spelling errors on lexical variation measures. Lexical varia- tion measures such as Type-Token Ration (TTR) and Yule’s K are often used for learner English analysis and assessment. When applied to learner English, however, they can be unreliable be- cause of the spelling errors appearing in it. Namely, they are, directly or indirectly, based on the counts of distinct word types, and spelling errors undesirably increase the number of distinct words. This paper examines the hypothesis that lexical variation measures become unstable in learner English because of spelling errors. Specifically, it tests the hypothesis on English learner corpora of three groups (middle school, high school, and college students). To be precise, it estimates the difference in TTR and Yule’s K caused by spelling errors, by calculating their val- ues before and after spelling errors are manually corrected. Furthermore, it examines the results theoretically and empirically to deepen the understanding of the influence of spelling errors on them. Title and Abstract in French Une exploration de l’influence des erreurs orthographiques sur les mesures de variation lexicale Cet article explore l’influence des erreurs orthographiques sur les mesures de variation lexicale. Les mesures de variation lexicale telles que le rapport type-occurrence (Type-Token Ratio; TTR) et le K de Yule sont souvent employ´ees pour analyser et ´evaluer l’anglais d’apprenants langue seconde. Lorsqu’on les applique a l’anglais d’apprenants, elles peuvent cependant s’av´erer peu fiables du fait des erreurs orthographiques que l’on y rencontre. De fait, elles se fondent directe- ment ou indirectement sur le d´ecompte des types de mots diff´rents, et les erreurs orthographiques augmentent artificiellement le nombre de mots diff´erents. Cet article examine l’hypoth`ese selon laquelle les mesures de variation lexicale deviennent instables sur l’anglais d’apprenants `a cause des erreurs orthographiques. Il teste cette hypoth`ese sur des corpus d’anglais d’apprenants de trois groupes (´el`eves de coll`ege, de lyc´ee, et ´etudiants `a l’universit´e). Plus pr´ecis´ement, il estime la diff´erence caus´ee par les erreurs orthographiques sur le TTR et le K de Yule en calculant leurs valeurs avant et apr`es correction manuelle des erreurs orthograpiques. Il examine de plus ces r´esultats th´eoriquement et empiriquement pour approfondir notre compr´ehension de l’influence des erreurs orthographiques sur ces mesures. 1 Introduction Vocabulary richness measures are often used for learner language analysis and assessment as in the work by Arnaud (1984), Attali and Burstein (2006), Ishikawa (2015), Gregori-Signes and Clavel-Arroitia (2015), and ˇSiˇskov´a (2012). Among a wide variety, Type-Token Ratio (TTR), which is defined as the This work is licensed under a Creative Commons Attribution 4.0 International License. License details: http:// creativecommons.org/licenses/by/4.0/. 2392 number of distinct words divided by the total number of words, is probably the most …

Content truncated to 500 words