Using speech recognition algorithms to improve the effectiveness of listening training in foreign language teaching

Background: Comprehension is a crucial element for accurate communication in foreign language learning. The outcomes of conventional teaching strategies are often poor due to their low level of preparation and responsiveness. Speech recognition technology, which provides immediate feedback, offers greater practicality and advantages; however, its integration requires careful consideration to ensure its effectiveness and benefits. Method: In this paper, we gathered speech signal datasets from 120 learners and divided them equally into two groups: a baseline group and a study group. The baseline group received instruction through a traditional technique, while the study group was taught using a proposed new technique named the Artificial Rabbit Optimized Hidden Markov Model (AR-HMM). We then conducted an analysis using the proposed technique to improve listening training outcomes in foreign English language learning. Result: First, we used nine questionnaires to examine the participants’ experience. We evaluated the performance of the proposed technique and compared it with an existing technique based on parameters such as accuracy (95%), involvement (low: 15%, medium: 59%, high: 93%), efficiency (90%), and time (14 h). Conclusion: Compared to the existing technique, our proposed technique demonstrates superior performance in improving the effectiveness of listening training in foreign English language learning.