China-ASEAN Academic Field Sentiment Analysis Model Based On CNN-LSTM

With the continuous deepening of China-ASEAN connectivity,the importance about the research of sentiment analysis in the China-ASEAN academic field is growing high.It not only provides strong support in the formulation of academic topics,but also provides strong data support for the government to analyze bilateral relations,predict opportunities and risks.However,the existing research on academic field sentiment analysis in China-ASEAN is insufficient,so based on CNN,RNN,LSTM,and other basic algorithms,combined with the grammatical characteristics of academic field articles,this paper puts forward A-AFSAM model to research sentiment analysis in China-ASEAN academic field,which not only have academic significance,but also have high application value.First of all,through the Internet crawler technology based on Python,crawling the content of academic papers in various fields about China-ASEAN,taking the abstract part of academic papers as the main analysis object;preprocessing the crawling corpus,data cleaning,data integration,data standardization,word segmentation,word embedding and other preprocessing operations.Then,the emotion classification model based on deep learning algorithm is constructed.According to the grammatical characteristics of academic articles and the algorithm of CNN,a C-AFSAM model is proposed;because the core of C-AFSAM cannot extract information in the time sequence of the corpus,a R-AFSAM model is proposed in combination with RNN algorithm;because the length of academic articles is long,the R-AFSAM model does not perform well in the long time sequence,so combining with the cell structure of LSTM,a M-AFSAM model is proposed.RAFSAM model is used to improve the memory of the model when processing sequence data.After the super parameters suitable for each model are determined by experiments,the results show that C-AFSAM model and M-AFSAM model have excellent performance in accuracy,precision,recall and F1 score.Considering the good performance and high efficiency of C-AFSAM model,and the best performance of M-AFSAM model,but the running efficiency of M-AFSAM model is low and takes a long time,the two models are integrated to obtain their respective advantages,a Combined-AFSAM model based on CNN and LSTM is proposed,which can reduce the dimension of data in advance by convolution and pooling operation in C-AFSAM model,extraction and other operations,and then the LSTM cells carry out deep data mining and other work.Through the experiment and analysis of the results,it is found that the combined model has a certain degree of improvement in accuracy and precision.And it makes a great progress in operating efficiency.Finally,it introduces the framework and structure of China-ASEAN ocean big data platform.Then integrates the text classification model of the academic field into the public opinion monitoring module of the academic field in big data platform.