| Aspect-level sentiment analysis is a kind of subtask of sentiment analysis research,which aims to study the sentiment tendency of the evaluation subject to an attribute or aspect of the evaluation target in a text.This task has a wide range of application scenarios in e-commerce websites and social media.This paper mainly considers the following problems in aspect level sentiment analysis:there is noise interference among sentiment words of multiple evaluation objects,and there is insufficient mining of fixed sentence patterns with emotional expression.A hierarchical graph neural network model is proposed and used to build a restaurant review system.The main work of this paper includes:a hierarchical graph neural network model is proposed.An end-to-end model for aspect level affective analysis task is designed.The model can avoid the noise interference caused by word position factor,and learn the syntactic structure which has indicative effect on affective polarity prediction.We designed and implemented an application system for restaurant reviews.Specifically,we apply the proposed sentiment analysis model to the restaurant review system,and the test results show that the model has achieved competitive results in performance and effect.At the same time,the function and stability of the whole system have reached the standard for online production environment. |
