方面级情感分析的研究与应用

Research and Application of Aspect-Based Sentiment Analysis

在早期的细粒度情感分析任务上,主要由人工进行特征选择,人为的工作量大,传统机器学习模型如支持向量机、朴素贝叶斯等被用于情感分类,但随着近几年神经网络的兴起,多种神经网络代替人为的工作,实验效率大大地提升,因为需要挖掘人们文本表达的观点的信息,本文就多种模型应用于方面级情感分析进行研究和分析,得出其中方面词和上下文之间的语义关系影响重大,针对该方面对模型进行应用,得到效果最好的模型。

In the early tasks of Aspect-based Sentiment Analysis, the feature selection was mainly carried out manually with a high human workload, and traditional machine learning models such as support vector machines and Naive Bayes model were used for sentiment classification. However, with the rise of neural networks in recent years, a variety of neural networks have replaced human work, and experimental efficiency has been greatly improved. Because of the need to tap into information about the views expressed in people’s texts, this paper studies and analyzes the application of various models in Aspect-based Sentiment Analysis, concluding that the semantic relationship between aspect words and context has a great influence, and that the models are applied to that aspect to obtain the best results.