Research on English Reading Comprehension Strategies Based on Natural Language Processing

Abstract

Abstract This study investigates strategies based on natural language processing to improve the efficiency and effectiveness of English reading comprehension by analyzing and optimizing reading comprehension models. We introduce the main framework of the English reading comprehension task, including the word embedding representation layer, the context encoding layer, the information interaction layer, and the output layer. The article utilizes the Word2vec model and decision tree ID3 algorithm to enhance the efficiency and accuracy of reading comprehension. The research data comprised 800 English articles, with 60% used for training and 40% used for testing. The articles were graded for difficulty and resources were pushed. It was found that the model proposed in this study performs better in difficulty grading (with a rating range of 0.797-0.926) and is more efficient in terms of time overhead (feature extraction time of about 10 milliseconds, and grading time of 1.7-2.6 seconds), compared to the traditional models such as BiLSTM and RNN. The pushed resources significantly correlate with learners’ self-assessed difficulty and actual performance. The NLP-based strategy proposed in the paper shows significant advantages in both efficiency and accuracy, and helps to improve English learners’ reading comprehension.