An Explainable Artificial Intelligence Framework for Improving English Vocabulary Learning Through Phonetic Similarity Between Familiar and Advanced Words for Language Learners

Learning advanced English vocabulary, particularly at the C۱ proficiency levels, can be challenging for language learners because unfamiliar words may involve complex and less familiar pronunciation patterns. This study proposes an explainable artificial intelligence framework for recommending advanced vocabulary based on phonetic similarity to words that learners are already familiar with at the A۱, A۲, B۱, and B۲ levels. In the proposed framework, the International Phonetic Alphabet (IPA) representations of words are converted into phoneme sequences, and phonetic correspondences between familiar and advanced words are identified using the Smith–Waterman alignment algorithm. Multiple phonetic features, including shared phoneme sequence length, phonetic similarity ratio, pronunciation coverage, the position of shared phonemes (prefix, middle, or suffix), and word-length similarity, are then combined to calculate a normalized similarity score and rank candidate C۱words. To enhance the interpretability of the recommendations, a Large Language Model (LLM) is subsequently used to generate learner-oriented explanations describing the shared pronunciation patterns and their locations within the advanced words. Experimental results show that the framework achieves a coverage rate of ۷۰% with an average similarity score of ۰.۷۴۸, indicating that a substantial proportion of familiar vocabulary items can be associated with suitable advanced words through phonetic similarity. The proposed framework provides an interpretable bridge between familiar and advanced vocabulary and offers a potential approach for integrating phonetic similarity and LLM-based explanations into intelligent English vocabulary and pronunciation learning systems.

کلیدواژه ها:

Artificial Intelligence ، English Vocabulary Learning ، Phonetic Similarity ، Pronunciation Learning ، Large Language Models ، Explainable AI