Teaching Practices to Enhance English Reading Comprehension Using Natural Language Processing Technology

With the emergence of large-scale pre-training models, natural language processing technology has made great progress. In this paper, natural language processing technology is combined with English reading comprehension to construct an English reading comprehension model. For the English reading comprehension model of Bert’s pre-training model, word combination is used, and dependent syntactic analysis and keyword co-occurrence are introduced to improve the encoding process of the model. The English reading questions and articles are vectorized, features are extracted, and text information and features are fused to generate the output of English reading comprehension answer intervals. The English reading comprehension system is built using the English reading comprehension model described in this paper as its technical core. In the testing experiments, the English reading comprehension model of this paper reaches as high as 88.36% and 85.05% in RACE-1 and RACE-2 datasets, and the corresponding accuracy also reaches 86.33% and 83.85% respectively, with a better performance than other baseline models. As for the educational practice of English reading comprehension, the posttest scores of the experimental class taught with the English reading comprehension system of this paper were higher than those of the pretest by 24.46, showing a significant difference (P=0.002<0.05), while the control class showed no significant change (P=0.15>0.05).