基于BiGCN和IAM的方面级情感分类模型

目前基于神经网络的方面级情感分类模型很少会考虑上下文单词与方面词之间的句法依存关系，可能会错误地将与方面词语法无关的上下文单词作为方面词的情感特征；另一方面大多数方法也忽略了上下文与方面词之间的交互信息。针对这两个问题，提出了基于双向图卷积网络（BiGCN）和交互注意力机制（IAM）的方面级情感分类模型（BiGCN-IAM），该模型在句法依存树上使用双向图卷积网络提取上下文单词和方面词之间的句法依存关系，然后使用掩码层得到特定的方面词表示；最后使用交互注意力机制学习上下文与方面词之间的交互信息，同时提取了上下文中的重要情感特征和方面词中对分类有贡献的特征。通过在五个公开数据集上的实验证明，该模型效果优于基线模型。

Aspect Level Sentiment Classification Model Based on BiGCN and IAM

At present, the aspect level sentiment classification model based on neural network seldom considers the related syntactic constraints and long-distance word dependency relation, and ignores the interaction information between the sentence and aspect words, and may mistakenly take context words irrelevant to aspect lexicon as sentimental features of aspect words. To solve two problems, this paper proposes an aspect level sentiment classification model（BiGCN-IAM） based on bidirectional graph convolutional neural network and interactive attention mechanism. The model uses bidirectional graph convolutional neural network to extract syntactic relation and dependency information between words on the syntactic dependency tree, and then uses the mask layer to get the specific aspect word representation. Finally, interactive attention mechanism is used to learn the interaction information between context and aspect words, and the important sentimental features in the context and the important classification features in the aspect words are extracted at the same time. Experiments on five public datasets show that the proposed model is better than the baseline model.