Abstract
Tourism is a form of cultural expression. The key step of fine-grained sentiment analysis of travel review evaluation text is opinion targets extraction. It aims to analyze travel review text and extract the opinion targets contained in it, and targets extraction’s accuracy rate directly affects the accuracy rate of sentiment analysis. Due to the outstanding performance of pre-training language models in the field of natural language processing in recent years, in order to improve the accuracy of opinion targets extraction, we propose a RoBERTa pre-training language model and Chinese pre-training word embedding vector, combined with BiLSTM (bidirectional long short-term memory) and CRF (conditional random field) opinion targets extraction model. The experimental results on the travel review data set of the Mafengwo travel software show that the extraction effect of this model is improved to varying degrees compared with other existing opinion targets extraction models based on deep learning.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
