Cross-domain sentiment classification via spectral feature alignment

By Sinno Jialin Pan, Xiaochuan Ni, Jian Sun, Qiang Yang, Zheng Chen. 2010

836 citations recorded by the source corpus.

Sentiment classification aims to automatically predict sentiment polarity (e.g., positive or negative) of users publishing sentiment data (e.g., reviews, blogs). Although traditional classification algorithms can be used to train sentiment classifiers from manually labeled text data, the labeling work can be time-consuming and expensive. Meanwhile, users often use some different words when they express sentiment in different domains. If we directly apply a classifier trained in one domain to other domains, the performance will be very low due to the differences between these domains. In this work, we develop a general solution to sentiment classification when we do not have any labels in a target domain but have some labeled data in a different domain, regarded as source domain. In this cross-domain sentiment classification setting, to bridge the gap between the domains, we propose a spectral feature alignment (SFA) algorithm to align domain-specific words from different domains into unified clusters, with the help of domain-independent words as a bridge. In this way, the clusters can be used to reduce the gap between domain-specific words of the two domains, which can be used to train sentiment classifiers in the target domain accurately. Compared to previous approaches, SFA can discover a robust representation for cross-domain data by fully exploiting the relationship between the domain-specific and domain-independent words via simultaneously co-clustering them in a common latent space. We perform extensive experiments on two real world datasets, and demonstrate that SFA significantly outperforms previous approaches to cross-domain sentiment classification.

Subfields: Artificial Intelligence

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. B. Liu
3. Soujanya Poria
4. Gabriella Csurka
5. Qiang Yang
6. Erik Cambria
7. Shuai Wang
8. Wouter M. Kouw

Citations in the corpus

• A Comprehensive Survey on Transfer Learning (2019, methodology): “SFA generally contains the following five steps.”
• Domain Adaptation for Visual Applications: A Comprehensive Survey (2017, background): “named entity recognition and opinion extraction across different text corpora [59, 161, 15, 276], multilanguage text classification [167, 275, 274], sentiment analysis [35, 100], WiFi-based localization [159], speech recognition across different speakers [132, 175], object recognition in images acqu”
• Learning with Augmented Features for Heterogeneous Domain Adaptation (2012, methodology): “Motivated by (Daum´e III, 2007; Pan et al., 2010), we also incorporate original features in this work and then augment any source and target domain samples x s ∈ R d s and x t ∈ R d t by using our newly proposed feature mapping functions φ s and φ t as follows: (1) After introducing P and Q , the…”
• WILDS: A Benchmark of in-the-Wild Distribution Shifts (2020, neutral): “…much smaller than modern datasets, e.g., in object recognition (Saenko et al., 2010; Gong et al., 2012), sentiment analysis (Blitzer et al., 2007; Pan et al., 2010), part-of-speech tagging (Marcus et al., 1993; Daumé III, 2007), flow cytometry (Blanchard et al., 2011; Muandet et al., 2013), or…”
• Forward and Backward Knowledge Transfer for Sentiment Classification (2019, background): “Our work is also closely related to transfer learning (TL) (Pan and Yang, 2010; Blitzer et al., 2007; Andreevskaia and Bergler, 2008; Bollegala et al., 2011; He et al., 2011; Li et al., 2012, 2013; Xia and Zong, 2011; Pan et al., 2010; Wu et al., 2009, 2017; Augenstein et al., 2018).”
• An introduction to domain adaptation and transfer learning (2018, methodology): “In order to find common components where the sets are not too dissimilar, the Maximum Mean Discrepancy is employed [154,156].”
• TrGraph: Cross-Network Transfer Learning via Common Signature Subgraphs (2015, background): “Transfer learning is a machine learning paradigm which aims at borrowing knowledge from a related domain to help build an accurate classifier in the target domain, where the labeled data in the target domain is very limited [9].”
• A survey of sentiment analysis in social media (2018, neutral): “[84] first proposed spectral feature alignment (SFA).”
• Transfer learning using computational intelligence: A survey (2015, background): “…statistics vary between domains [24, 30-32] or by learning a feature space projection in which a distribution divergence statistic is minimized [33-35]; (B) Latent feature approaches aim to construct new features by analyzing large amounts of unlabeled source and target domain data [25, 36-42].”
• Towards multidomain and multilingual abusive language detection: a survey (2021, methodology): “Several methods have been proposed by studies in more mature areas, such as sentiment analysis [35, 95, 151].”
• Transfer Learning with Graph Co-Regularization (2012, neutral): “2011), sentiment analysis (Pan et al. 2010a), image classification (Zhu et al.”
• Bilingual Co-Training for Sentiment Classification of Chinese Product Reviews (2011, methodology): “2009), and spectral feature alignment (Pan et al. 2010).”
• A Survey on federated learning* (2020, background): “Transfer learning [17] enables to move the knowledge of one domain (i.”