Improving English Foreign Language (EFL) Performance using Artificial Intelligence in Vocational Education and Training (VET)

Abstract

Internationalisation is one of the strategies for improving the technical qualifications and employability of trainers in initial and continuing vocational education and training. It is based on the full development of linguistic competence in a foreign language such as English, which is influenced by various factors, including affective factors. Currently, one resource for detecting poor performance in English is artificial intelligence to the extent that it can predict academic performance. This research aims to predict performance in English as a foreign language based on affective variables such as willingness to communicate orally in English, self-efficacy and English language anxiety. The experimental result shows that the prediction model trained with a decision tree algorithm (J48) provides the best data for predicting performance in English in terms of accuracy = 0.74, precision = 0.70, recall = 0.678 and F-score = 0.68. Analysing the influence of the variables and eliminating the data for the affective variable willingness to communicate orally in English yields the best accuracy = 0.76. This finding has relevant practical implications for the early identification of underachievement in English and for personalising educational interventions to improve learning and performance in English as a foreign language among vocational education and training students.