Intelligent scoring method for English articles leveraging large language models

Traditional manual scoring of English essays faces subjectivity and low efficiency, while existing automated essay scoring (AES) systems suffer from limited genre adaptability, weak scoring-feedback linkage, and poor robustness.To address these challenges, this study proposes a BERT and Zhipu AI model-based automatic scoring algorithm for different English genre essays (BERT-AI-ASEG), integrating 'overall scoring feature scoring feedback output'.Furthermore, an English article intelligent scoring algorithm integrating multiple large language models (EAIS-IMLLM) is developed to enhance robustness.Experiments on the ASAP++ dataset show that the Spearman's rank correlation coefficient and quadratic weighted kappa reach 0.87 and 0.85, improving by 4.82% and 6.25% over comparison algorithms.The proposed algorithm also achieves an evaluative feedback correlation of 0.89 and a multi-genre adaptation index of 0.91 at 150 iterations.These results indicate that the method is effective in improving accuracy, reliability, and teaching feedback effect for large-scale English writing assessments.