Design And Implementation Of Product Public Opinion Analysis System Based On Hierarchical Sentiment Analysis

With the rapid development of online social networking,shopping and other platforms,the Internet is filled with a lot of user’s thinking,emotional tendencies and other content.How to dig out valuable information reasonably and effectively through the public opinion analysis system has attracted more and more attention of scientific research scholars and related companies,and the active development of sentiment analysis technology as the core of public opinion analysis also rely on this.In recent years,the research on sentiment analysis technology has made great progress,but there are still certain limitations.Specifically,dictionary-based sentiment analysis methods have problems such as high labor costs and poor transferability.Document-level and sentence-level sentiment analysis methods based on machine learning have coarse granularity,and it is difficult to effectively aggregate different sentiment views in a large amount of data.At the time when data is becoming more and more complex,the flat spreading of sentiment categories in mainstream aspect sentiment analysis methods limits the depth of analysis,and it is easy to overlook the inherent correlation between different sentiment views.Under the above background,this paper drawing on the idea of hierarchical classification studies and implements the Hierarchical Category Sentiment Analysis(HCSA)algorithm that can refine the sentiment views into layers.And this paper designs and implements a product public opinion analysis system based on hierarchical sentiment analysis.The main design and implementation work of this paper are as follows:(1)This paper designs and implements the Hierarchical Category Sentiment Analysis(HCSA)algorithm.First,in view of the shallow depth and weak scalability of mainstream methods,this paper proposes a hierarchical sentiment analysis method with hierarchical and detailed sentiment view.Secondly,on this basis,this paper designs the Hierarchical Category Sentiment Analysis algorithm model with a multi-level structure using gate convolution,and combines the attention mechanism to improve the performance.Then,aiming at the small sample problem of deep sentiment category,this paper designs the hierarchical dependency prediction method and the hierarchical reward and punishment loss function which uses the dependency relationship between levels to alleviate the problem.Finally,in the data set composed of real data,this paper carries out the simulation experiments of the hierarchical sentiment analysis task of the proposed algorithm and different improvement points.Compares a variety of related algorithms,the main indicators of the algorithm are improved,and the effectiveness of the improvement points is verified.(2)This paper takes the proposed hierarchical category sentiment analysis(HCSA)algorithm as the core,and uses the Django framework to design and implement a product public opinion analysis system based on hierarchical sentiment analysis.The performance has been tested,the test results verify the effectiveness and reliability of the function,and system can meet the needs of product public opinion analysis under the type of hierarchical sentimental views proposed in this paper.