An improved aspect-category sentiment analysis model for text sentiment analysis based on RoBERTa
- PDF / 4,261,323 Bytes
- 12 Pages / 595.224 x 790.955 pts Page_size
- 13 Downloads / 359 Views
An improved aspect-category sentiment analysis model for text sentiment analysis based on RoBERTa Wenxiong Liao1 · Bi Zeng1 · Xiuwen Yin2
· Pengfei Wei1
Accepted: 19 September 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract The aspect-category sentiment analysis can provide more and deeper information than the document-level sentiment analysis, because it aims to predict the sentiment polarities of different aspect categories in the same text. The main challenge of aspect-category sentiment analysis is that different aspect categories may present different polarities in the same text. Previous studies combine the Long Short-Term Memory (LSTM) and attention mechanism to predict the sentiment polarity of the given aspect category, but the LSTM-based methods are not really bidirectional text feature extraction methods. In this paper, we propose a multi-task aspect-category sentiment analysis model based on RoBERTa (Robustly Optimized BERT Pre-training Approach). Treating each aspect category as a subtask, we employ the RoBERTa based on deep bidirectional Transformer to extract features from both text and aspect tokens, and apply the cross-attention mechanism to guide the model to focus on the features most relevant to the given aspect category. According to the experimental results, the proposed model outperforms other models for comparison in aspect-category sentiment analysis. Furthermore, the influencing factors of our proposed model are also analyzed. Keywords Sentiment analysis · Attention mechanism · Convolutional neural network · Text classification
1 Introduction With the development of the internet, social media is growing into one of the main platforms for people to share information and their personal opinions in daily life [1]. As an important application of social information analysis, sentiment analysis [2] has aroused wide concern. According to the results of sentiment analysis, different people
Xiuwen Yin
[email protected] Wenxiong Liao [email protected] Bi Zeng [email protected] Pengfei Wei [email protected] 1
School of Computers, Guangdong University of Technology, Guangzhou, 510006, China
2
School of Automation, Guangdong University of Technology, Guangzhou, 510006, China
may have different opinions and emotional tendencies on the same issue, which can provide powerful functions for competition analysis and market analysis. As the fine-grained sentiment analysis [3], the aspect-category sentiment analysis aims to provide deeper and more comprehensive results. For example, in the sentence “the food in this restaurant is delicious and the staff is very friendly, but the price is expensive”, the polarity is positive if the aspect category is food or service, but the polarity is negative if the aspect category is price. Aspect-category sentiment analysis belongs to the scope of text classification, which considers the information of both text and aspect categories. Aspect-category sentiment analysis is a kind of aspect-based sentiment analysis [4]. C
Data Loading...
