Research on Text Classification Based on Convolutional Neural Network

Abstract

Text classification is one of the research hotspots in the field of Natural Language Processing (NLP). In this paper, TextCNN model based on Convolutional Neural Network (CNN) is used for classification; the classified corpus is selected from the text extracted from the electronic instruction manual. During the experiment, the text was preprocessed at first, then the processed text was converted intoword vector formatto generate data sets, which were finally input into TextCNN for training. In order to verify the influence of TF-IDF weighted word vectors on training results, data sets made by weighted and unweighted word vectors were used in the comparison experiment to conduct classification model training and to calculate the final accuracy of the model, it is concluded that the classification accuracy can be improved by using weighted word vectors.