A hierarchical lstm model with multiple features for sentiment analysis of sina weibo texts
Shumin Shi, Meng Ya Zhao, Jun Tao Guan, Yaxuan Li, Heyan Huang · 2017
Sentiment analysis has long been a hot topic in natural language processing. With the development of social network, sentiment analysis on social media such as Facebook, Twitter and Weibo becomes a new trend in recent years. Many different methods have been proposed for sentiment analysis, including traditional methods (SVM and NB) and deep learning methods (RNN and CNN). In addition, the latter always outperform the former. However, most of existing methods only focus on local text information and ignore the user personality and content characteristics. In this paper, we propose an improved LSTM model with considering the user-based features and content-based features. We first analysis the training dataset to extract artificial features which consists of user-based and content-based. Then we construct a hierarchical LSTM model, named LSTM-MF (a hierarchical LSTM model with multiple features), and introduce the features into the model to generate sentence and document representations. The experimental results show that our model achieves significant and consistent improvements compared to all state-of-the-art methods.