Design And Implementation Of Text Classifier Based On Convolution Neural Network

With the rapid development of the Internet and the continuous expansion of information on the Internet,in the face of a vast ocean of information,people are faced with enormous information disasters while accessing abundant information resources.In order to provide efficient and accurate information services,a series of information processing technologies have been produced.Text classification is an important task in modern information processing and an important part of text mining.First of all,this paper describes the general flow of text classification tasks,including:data preprocessing,word segmentation,feature selection,classification algorithm research,and specifically analyzes the advantages and disadvantages of traditional classification algorithms.In the experiment,used the standard long text data set of Shanghai Fudan University.Compared with the short text data in the actual project,it was found that the traditional algorithm has a poor effect on short text data sets.Therefore,the deep learning method based on convolutional neural network used in this paper is derived.Secondly,this paper elaborates the related theory knowledge of convolutional neural network.Based on this,it designs and implements a text classifier based on convolutional neural network,which improves the classification accuracy of e-commerce short text data sets.Then,used random and Skip-gram two word vector representation methods to further design a two-channel convolutional neural network classification model.This model can be used to better complete the text based on the original input data.The degree of present and degree of reduction.Experiments show that compared with traditional machine learning methods,including support vector machine and naive Bayes,the improved dual-channel convolutional neural network model proposed in this paper improves the classification accuracy of e-commerce short text data sets.Third,in order to use the above research results in the project,this paper completes the system's requirements analysis,outline design,detailed design,and system testing according to the software engineering development specification method,and implements a text classification system based on convolutional neural network.This text classification technology system can predict the type of short text input by the user and meet the desired effect.Finally,on the basis of summing up the full text,the future direction of the work is put forward.