Computer Science · Artificial Intelligence
Retracted: An Investigation on Sentiment Analysis merged with CNN and Deep BI-GRU Model
See on the mapAbstract
Convolutional neural networks (CNN) and recurrent neural networks (RNN) have been utilised in message sentiment analysis, and both have produced promising results. CNN overlooks the coherent semantic information between words despite its ability to separate nearby information between successive articulations of a sentence. Many of the current sentiment analysis methods are seeking to extract the complicated semantic also, linguistic data from the message, however they are not especially applicable to the examination of understudy opinions. For the study of student message sentiment, the CNN-BiGRU-AT model—often referred to as a novel convolutional neural network with the bidirectional gated recurrent unit and an attention mechanism—is developed. The removal of n-gram information from each sentence is subsequently accomplished using a convolutional neural network (CNN). The sentences are then consecutively condensed through the bidirectional gated intermittent unit to dispose of the message's significant semantic data parts (BiGRU). The CNN-BiGRU model is then given a thinking mechanism, and by including the attention score, different learning loads are introduced to the model. The disclosures display that its structure's accuracy rate and audit rate dependably outperform 0.9, and the F1 regard isn't lower than 0.8. These figures are superior to those of various models.