Learning Semantic Representations of Users and Products for Document Level Sentiment Classification

By Duyu Tang, Bing Qin, Ting-Wei Liu. Annual Meeting of the Association for Computational Linguistics, 2015

363 citations recorded by the source corpus.

Neural network methods have achieved promising results for sentiment classification of text. However, these models only use semantics of texts, while ignoring users who express the sentiment and products which are evaluated, both of which have great influences on interpreting the sentiment of text. In this paper, we address this issue by incorporating userand productlevel information into a neural network approach for document level sentiment classification. Users and products are modeled using vector space models, the representations of which capture important global clues such as individual preferences of users or overall qualities of products. Such global evidence in turn facilitates embedding learning procedure at document level, yielding better text representations. By combining evidence at user-, productand documentlevel in a unified neural framework, the proposed model achieves state-of-the-art performances on IMDB and Yelp datasets1.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/6ee62c42bb00d8ddbff898ca82cf1c7699ec79e9

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Bing Liu
2. Erik Cambria
3. Zhiyuan Liu
4. Maosong Sun
5. B. Liu
6. Ting-Wei Liu
7. Shuai Wang
8. Duyu Tang

Citations in the corpus

• Interpretable Convolutional Neural Networks with Dual Local and Global Attention for Review Rating Prediction (2017, methodology): “Most networks use word embeddings [12, 22, 24, 25, 28] as the input layer, while [5, 7, 29] used character embeddings, especially [5] uses very deep (29 layers) CNNs on character embeddings and shows impressive results on text classi cation, including the Amazon reviews.”
• A review of natural language processing techniques for opinion mining systems (2017, background): “For product reviews, user and product information are adopted to capture clues about an overall sentiment polarity [15].”
• Document Modeling with Gated Recurrent Neural Network for Sentiment Classification (2015, methodology): “Specifically, we try CNN with multiple convolutional filters of different widths (Tang et al., 2015) to produce sentence representation.”
• Deep learning for sentiment analysis: A survey (2018, background): “Tang et al. 40 applied user representations and product representations in review classification.”
• Neural Sentiment Classification with User and Product Attention (2016, methodology): “For all baseline methods above, we report the re-sults in (Tang et al., 2015b) since we use the same datasets.”
• Learning to Spot and Refactor Inconsistent Method Names (2019, background): “Recent literature has provided evidence that paragraph vector outperforms other state-of-the-art techniques [35], [36] for text representations [32], [37], and can effectively capture semantic similarities among words and sentences [38]–[42].”
• Sentiment analysis using deep learning architectures: a review (2019, background): “…data (Wang et al. 2016) Capture the local semantics of n-grams of various granularities, which are proven powerful for sentiment classification (Tang et al. 2015) Identify fixed length phrases in consumer review domain (Do et al. 2019) Capture the most important n-grams feature of a sentence…”
• Sentiment Classification by Leveraging the Shared Knowledge from a Sequence of Domains (2019, background): “Many neural network based sentiment classification techniques have been reported [4,27,32,41,48].”
• Recent advances in deep learning based dialogue systems: a systematic survey (2021, methodology): “Yelp It consists of customer reviews from Yelp Dataset Challenge Response Retrieval Tang et al. (2015) Debates The dataset consists of debates on Congerssional bills.”
• A Survey on Context Learning (2017, neutral): “[72] shows how to utilize user and product context information when dealing with sentiment classification task of text.”
• Like It or Not (2016, methodology): “Recently, Tang et al. [2015a] used deep learning to learn semantic representations of user and products, whereas Tang et al. [2015b] used deep learning for review prediction.”