Personalized English Learning Resource Recommendation using Attention Long Short-Term Memory

English language learning has changed with the incorporation of digital technologies, and personalized recommendation systems have taken center stage to enhance the learning process. Classical recommendation systems typically use rule-based algorithms or collaborative filtering, which do not map educational resources precisely to the individual needs, preferences, and learning progress of the learner. These approaches frequently grapple with content mismatch, inflexibility, and limited capacity to support dynamic learning behavior, resulting in decreased learner engagement and ineffective learning experiences. To overcome these challenges, this study introduces a new English teaching resource recommendation mechanism based on Attention LSTM (Long Short-Term Memory) networks. The model utilizes the advantages of LSTM networks to handle sequence data and incorporates an attention mechanism to pay specific attention to the most important learning factors, including proficiency level, learning gaps, and changing preferences. By giving priority to important content, the model provides more personalized and efficient recommendations. Experimental findings proves the proposed LSTM model by attaining accuracy of 0.98, precision of 0.97, recall of 0.965 and F1-scor of 0.972. Comparative analysis with conventional recommendation methods reveals that the model provides increased accuracy, learner satisfaction, and learning engagement. The findings highlight the potential of Attention LSTM networks to enhance personalized English language learning by overcoming the drawbacks of current methods and providing an adaptive, data-driven strategy to resource recommendation.