A Study on Enhancing English Reading Comprehension Using Natural Language Processing Technology

Le résumé fourni par la source

In order to solve the problem of low efficiency and accuracy of traditional English software’s word division, the research of using natural language processing technology to improve English reading comprehension is proposed. First, we construct the structure level of the English translation scoring system, which includes translation data collection module, information feature extraction module, analysis model construction module and result feedback scoring module; we establish the language model of the English translation scoring system, use the model to statistically determine the probability distributions of specific sentence sequences or word sequences of the translations, and carry out the extraction of the information features of the user’s English translation document and the training set of the translation; based on the feature extraction results, we calculate the feature keyword similarity, and use particle swarm optimization to optimize the keywords. extracted; the feature keyword similarity is calculated according to the feature extraction results, and the particle swarm optimized BP network is used for the fitting calculation, and a hybrid GRU-CRF network segmentation method is also proposed to accelerate the model training efficiency and segmentation accuracy. Analyzed by simulation. The experimental results show that the accuracy rate (Ldis +Lgen ) is up to 0.853 and the accuracy rate (Ldis ) is up to 0.828 when the data dimension is 640. Conclusion: The results show that the proposed hybrid GRU-CRF network segmentation method has faster convergence speed and better model performance.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.