CNN-LSTM-based Method for Automatic Recognition of Emotional Elements in Big Data's Comment Text on Social Media
To accurately identify the sentiment of social media comment texts and assist in guiding public negative emotions,an automatic identification method of emotional element in social media big data comment texts based on CNN-LSTM is proposed.Through the classification of social media big data,comment texts are converted into numerical lists using tokens with dictiona-ry functions.Combined with word embedding technology,a vector list is obtained to complete the preprocessing of vector con-version for social media big data.The vector list obtained through preprocessing is input into the CNN network to obtain the fi-nal local feature values of emotional element in comment texts.These values are then transmitted to the LSTM,where they are adjusted through the forget gate,input gate,and output gate to obtain the feature representation results of emotional element in comment texts.After classification by the Softmax classifier,automatic identification of emotional element is achieved.Experi-mental results show that this method can effectively complete the preprocessing of experimental data,label positive and nega-tive emotional element in the form of text and tags,and accurately identify emotional element,indirectly reflecting social issues and demonstrating strong applicability.
social media datacomment textemotional elementvector listCNN-LSTMautomatic recognition
万方数据
维普
