Analysis Of Emotional Tendency Of News In ASEAN Countries Based On Deep Learning

With the rapid development of the Internet era,more and more attention has been paid to emotional tendency classification of network information.As China's "one belt,one road" development and the formation of the China-ASEAN Cooperation Center for marine development,it is of great significance to study the data value of ASEAN countries in related fields.As a medium of information transmission,news report has universality and can fully reflect the current situation of public opinion.Therefore,the analysis and Research on the emotional tendency of ASEAN news has certain application value.This paper mainly analyzes the emotional tendency of ASEAN news based on deep learning method.First of all,we use crawler method to obtain news data of ASEAN countries to build news database,and use more valuable mainstream media news data.Then,we preprocess the original news data according to certain rules,such as emotional annotation and text cleaning.Then we use word embedding technology Word2Vec algorithm to quantify the text.Finally,we build a network classification model to judge the emotional tendency.The deep learning models of CNN-ANET(ASEAN News Emotional Tendency)and LSTM-ANET are constructed based on the self-constructed ASEAN news data set,and the experimental results on the whole are higher than the emotional classification results based on traditional machine learning methods(Naive Bayes,k-Nearest Neighbors,Support Vector Machine).Considering the influence of context information association on the classification results in the emotional discrimination of news texts,this paper uses the long short-term memory network model combined with the particularity of news texts,and on this basis,introduces the attention mechanism to propose a long short-term memory network based on attention mechanism(Bi-LSTMAttention-ANET)model,which has significantly improved the classification accuracy.However,there is over fitting phenomenon.In order to solve this problem,an attention mechanism based on long short-term memory network emotional tendency analysis(ATA-BLSTM-ANET)model is proposed,which can increase the disturbance in the standardization of word embedding,enhance the robustness of the model,prevent overfitting phenomenon and improve the accuracy of the classifier.Finally,the ATA-BLSTM-ANET emotional tendency model proposed in this paper is applied to the actual project.The ocean news module of China ASEAN ocean big data information service platform shows the news tendency by classification,which provides strong support for the analysis and research of ASEAN countries' news related fields.