Intelligent Linguistic Analysis for Automated Assessment and Feedback in Language Education

The ongoing digital transformation from traditional pedagogy to intelligent language education requires a new integration of corpus analysis, sentiment analysis, and regression-based assessment models. It has sparked what has been dubbed as an “automated feedback revolution” and given rise to an “intelligent assessment paradigm” and “data-driven pedagogy.” The aim of this study was to describe and explore learners’ and teachers’ perceptions of their current level of competence in automated feedback systems and identify distinct challenges in adoption. Combining data from classroom transcripts, the Intelligent Language Education Survey, with the AntConc corpus analysis, we employ regression modelling to estimate predictive relationships in linguistic performance in language education in Uzbekistan. The data were analyzed by sentiment analysis techniques. The most interesting finding is that after controlling for demographic variables, sentiment scores become significant. The highest predictive accuracy is in writing and interactive speaking. The preferred feedback mechanisms were those which were adaptive and which safeguard learner autonomy more effectively; teachers made many references to digital platforms circulating feedback, providing guidance to learners in various ways. We conclude by considering the pedagogical implications of this research.

Ҳали таржима қилинмаган