Design And Implementation Of An English Online Writing And Reviewing System Based On NODEJS

This paper discusses the English text classification and writing system based on deep learning and NODEJS.Recent years,with the development of deep learning,it has made good progress in text processing,speech processing and other areas.And it also has shown a far-reaching prospect in English composition and writing classification.This paper uses NODEJS,an advanced technique originated from Goggle,established an online English writing system,which has prominent significance for online English education purpose.Through the method of deep learning to classify the text of English composition,the system has improved the original classification model of English text,and has a better accuracy rate.And it is of certain reference significance for the text classification based on deep learning.The first chapter of the analysis introduces the research background and the current application situation.Through analyzing the current situation of English text classification and English writing system based on deep learning,it gives the necessity of the system design to solve the existing problems,and then the necessity of the system research is expounded.The second chapter analyzes the requirements of the system in detail,introduces the system functions and research contents of the system,and designs the overall architecture of the system.The third chapter exerts efforts for researching and designing the English text classification module based on deep learning.This part classifies the text by the convolution neural network(CNN),which is a widely used deep learning algorithm,and improves the neural network model by adding the DROPOUT layer,the improved pool layer and the text preview for the aim to better train the model,which improves the accuracy of classification of neural networks.It compares the effect of text preprocessing on the training results,and the influence of the new pooling method and the universal maximum pool method on the text classification results.In addition,it explores and compares the influence of different number of iterations and different text word vector dimensions on the training results,and then obtained the optimal text classification neural network model.The fourth chapter introduces the realization of the front end of the English text automatic correction based on the BK-tree algorithm.This module corrects English text through the BK-tree of the data structure,and increases the speed and accuracy of the error correction in English text through the establishment of a dictionary and a wrong-word based library.