Interactive Attention-Based Convolutional GRU for Aspect Level Sentiment Analysis

DOI

10.2991/hcis.k.210704.002How to use a DOI?

Keywords

Sentiment classification; convolutional neural network; gated recurrent units; attention mechanism

Abstract

Aspect level sentiment analysis aims at identifying sentiment polarity towards specific aspect terms in a given sentence. Most methods based on deep learning integrate Recurrent Neural Network (RNN) and its variants with the attention mechanism to model the influence of different context words on sentiment polarity. In recent research, Convolutional Neural Network (CNN) and gating mechanism are introduced to obtain complex semantic representation. However, existing methods have not realized the importance of sufficiently combining the sequence modeling ability of RNN with the high-dimensional feature extraction ability of CNN. Targeting this problem, we propose a novel solution named Interactive Attention-based Convolutional Bidirectional Gated Recurrent Unit (IAC-GRU). IAC-GRU not only incorporates the sequence feature extracted by Bi-GRU into CNN to accurately predict the sentiment polarity, but also models the target and the context words separately and learns mutual influence between them. Additionally, we also incorporate the position information and Part-of-Speech (POS) information as prior knowledge into the embedding layer. The experimental results on SemEval2014 datasets show the effectiveness of our proposed model.

Copyright

© 2021 The Authors. Publishing services by Atlantis Press International B.V.

Open Access

This is an open access article distributed under the CC BY-NC 4.0 license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)
View full text (HTML)

TY  - JOUR
AU  - Lisha Chen
AU  - Tianrui Li
AU  - Huaishao Luo
AU  - Chengfeng Yin
PY  - 2021
DA  - 2021/07/20
TI  - Interactive Attention-Based Convolutional GRU for Aspect Level Sentiment Analysis
JO  - Human-Centric Intelligent Systems
SP  - 25
EP  - 31
VL  - 1
IS  - 1-2
SN  - 2667-1336
UR  - https://doi.org/10.2991/hcis.k.210704.002
DO  - 10.2991/hcis.k.210704.002
ID  - Chen2021
ER  -