In this paper, oriented to the needs of comprehensive literacy assessment for English majors, artificial intelligence technology is used as the main means to propose a comprehensive literacy assessment system for students based on neurocognitive diagnosis. The main modules of the system are user information management, import and weighted summation of grades, certificate statistics, certificate management and query, and students’ comprehensive literacy assessment and statistics. The literacy assessment function embeds the one-hot coding of students and exercises on the premise of constructing the prerequisite relationship of knowledge points in the modified Q-matrix, and uses the hidden layer of neural network to simulate the interaction between the two. As a result, a neurocognitive diagnostic model based on the modified Q matrix is formed, which realizes the diagnosis of learners’ mastery of knowledge points from the neurocognitive level. After processing the students’ comprehensive literacy data using principal component analysis, the PLS-SVM algorithm was designed to extract the features of the input data, and the extracted features were used as the SVM to build a prediction model of students’ comprehensive literacy. Selecting the experimental samples and setting the comprehensive literacy evaluation dimensions, the comprehensive literacy prediction model for English majors is built within this framework, and the prediction error rate of the model is <0.600.
