A personalized recommendation algorithm for English exercises incorporating fuzzy cognitive models and multiple attention mechanisms

In the era of digital education, the rapid growth and disordered distribution of learning resources present new challenges for online learning. However, most of the exercise recommendation systems lack targeted guidance and personalization. In response to the above problems in the recommendation of English reading, writing and translation exercises in higher education, this paper proposes an algorithm that combines improved fuzzy cognitive diagnosis model and multi-head attention mechanism to predict students' performance on future recommended exercises. Firstly, considering three factors: student response time, continuous answer time, and knowledge gap parameters, the cognitive attributes of English knowledge are determined from both subjective and objective aspects, so as to construct an improved fuzzy cognitive diagnosis model. Secondly, the multi-head attention mechanism is used to allocate appropriate attention weights to students' current knowledge status based on their past knowledge status. Finally, the performance of the algorithm is proved through experimental data.