Research on new media semantic analysis method based on GRU
Traditional sentiment analysis largely relies on the quality of sentiment dictionaries and feature extraction,which greatly limits the correct expression of media sentiment semantics.It is not until deep learning that semantic analysis of new media is advanced.The article constructs a new media semantic fast analysis model based on GRU,data training is conducted on relevant models and applied to practical cases.After comparison with traditional models,the following conclusions are drawn:Compared with SVM and LSTM models,the GRU model has stronger analytical ability and higher classification accuracy,up to 0.909;The GRU model has better emotional polarity judgment ability,but slightly lacks in neutral word judgment;The GRU model has stronger predictive ability.
NETL
NSTL
万方数据
