Using AI Approach to Explore Vietnamese ESL Students’ Perceptions on Integrations of Local Culture into English Language Teaching

This study aims to explore perceptions of English as a Second Language (ESL) students on integrations of local culture into English language teaching using Artificial Intelligence approaches. Research samples included 511 ESL students of the English Faculty of four public universities in Vietnam’s Mekong Delta region. The input factor dimensions comprise demographics, level of local culture integration, facilities, and curriculum-related factors. The output factor was necessities for local culture–English-teaching integration. Two supervised machine learning algorithms, Decision Tree (DT) and Support Vector Machine (SVM), were applied with oversampling to address data imbalance issues. Results indicated that the oversampling case achieved the highest performance. The research shows that the DT model was slightly better than the SVM with an accuracy of 97% and AUC of 98%. Feature importance analysis identified curriculum, facilities, and students’ hometown as key predictors. The findings provided empirical evidence to support data-informed curriculum reform in culturally responsive English language teaching. This study also develops a novel method to explore students’ perceptions and offers practical suggestions for improving academic quality. This study is expected to enhance institutional student recruitment while contributing to the region’s sustainable development and the broader goal of preserving intangible cultural heritage in Vietnam’s Mekong Delta.