In recent years,machine learning and deep learning technologies have made remarkable achievements in the Natural Language Processing(NLP).Text classification is one of the fundamental task in NLP.Text classification refers to classify text into corresponding categories.In this study,Public data sets were used in the experimental and Three relevant models were proposed.One of them is based on an improved term frequency statistical method PTF-IDF(Promoted Term Frequency Inverse Document Frequency)algorithm.Factors,such as intra-class dispersion and inter-class concentration,were added to the original TF-IDF algorithm to improve the performance of the model.The other two are based on deep learning technologies,which named TS-LSTM(Title Semantic-Long Short-Term Memory)network and MA(Multi-Attention)network.The former adds topic semantic information to the classification model and improves the LSTM network structure;the latter uses the MA network to aggregate the context information of the text,and simplifies the structure of the model to reduce model parameters.Both of them improve the accuracy of the classification task. |