Insights from Russian second language readability classification: complexity-dependent training requirements, and feature evaluation of multiple categories

Abstract

I investigate Russian second language readability assessment using a machine-learning approach with a range of lexical, morphological, syntactic, and discourse features. Testing the model with a new collection of Russian L2 readability corpora achieves an F-score of 0.671 and adjacent accuracy 0.919 on a 6-level classification task. Information gain and feature subset evaluation shows that morphological features are collectively the most informative. Learning curves for binary classifiers reveal that fewer training data are needed to distinguish between beginning reading levels than are needed to distinguish between intermediate reading levels.

Relevant Parameters for the Classification of Reading Books Depending on the Degree of Textual Readability in Primary and Compulsory Secondary Education (CSE) Students

“…As a prospective line of research, work will continue on this matter to incorporate literary and press texts to strengthen the model, which will be focused upon a wider age rank. In line with Reynolds' proposal [14], the results find their application in the educational sphere, to contribute to develop the selection of texts made in virtue of their characteristics, to foster a better acquisition of reading competence and to emphasize comprehension.…”

“…The complexity and difficulty of texts depend on their vocabulary and sentence length [14]. To measure the lexical complexity of a text, four parameters have been taken into account: lexical density, lexical diversity, lexical wealth, and average frequency per word, since these are the most employed parameters in diverse researches [57].…”

“…In this way, comprehensive reading results crucial for both academic and professional success, since it is a vehicle for the social inclusion in the contemporaneous world [10]. In this author's opinion, the access to equality of chances is regulated by the individual's reading competence, in such a way that the ability to exercise some of the social rights requires to previously incorporate this ability [10] In this same sense, Reynolds has indicated that the command of the reading skill is basic in the modern world [14].…”

An editable learner model for text recommendation for language learning

“…According to some studies in second language acquisition, vocabulary knowledge and grammatical knowledge have comparable effects on reading comprehension (Shiotsu & Weir, 2007;van Gelderen, Schoonen, Stoel, de Glopper & Hulstijn, 2007). However, others suggested vocabulary knowledge to be the better predictor (Brisbois, 1995;Haynes & Carr, 1990;Mecartty, 2000) and reported a significant correlation between vocabulary difficulty and text difficulty (François & Fairon, 2012;Heilman, Collins-Thompson, Callan & Eskenazi, 2007;Reynolds, 2016). While a number of computer-assisted language learning systems have addressed grammatical knowledge (e.g.…”

Textometr: an online tool for automated complexity level assessment of texts for Russian language learners

“…С развитием технологий автоматической обработки естественного языка в мировой научной практике появляются работы, посвященные возможностям автоматизации процесса оценки доступности текста (DuBay, 2004). История разработки автоматизированной оценки сложности русских текстов для преподавания иностранной аудитории пока не столь богата, как, например, для англоязычных текстов, однако все же содержит несколько пионерских работ (Karpov et al, 2014;Reynolds, 2016;Sharoff et al, 2008), на которые опирается данное исследование.…”