In recent years, Aspect-based Sentiment Analysis (ABSA) has been extensively researched in various languages because it aims to detect the sentiment of each aspect of the text. The ABSA problem can be divided into three subtasks as follow: the aspect detection, Opinion Target Expression (OTE) and Sentiment Polarity. In this paper, we present a Convolutional Neural Network architecture for the aspect detection for Vietnamese. The aspect detection is to aim to identify of the entity E and attribute A pairs expressed in the text (Pontiki et al., 2016). The experimental results show the superiority of our model over the winning systems on datasets of the VLSP 2018 challenge for aspect detection task. Our method achieves the F1 score of 80.40% for the restaurant domain and 69.25% for the hotel domain.
Paper
Full text
Deep Learning for Aspect Detection on Vietnamese Reviews
Semantic Scholar · Computer Science · 2018
Abstract
In recent years, Aspect-based Sentiment Analysis (ABSA) has been extensively researched in various languages because it aims to detect the sentiment of each aspect of the text. The ABSA problem can be divided into three subtasks as follow: the aspect detection, Opinion Target Expression (OTE) and Sentiment Polarity. In this paper, we present a Convolutional Neural Network architecture for the aspect detection for Vietnamese. The aspect detection is to aim to identify of the entity E and attribute A pairs expressed in the text (Pontiki et al., 2016). The experimental results show the superiority of our model over the winning systems on datasets of the VLSP 2018 challenge for aspect detection task. Our method achieves the F1 score of 80.40% for the restaurant domain and 69.25% for the hotel domain.