A Mixed Model for Cross Lingual Opinion Analysis

Gui, Lin; Xu, Ruifeng; Xu, Jun; Li, Yuan; Yao, Yong; Zhou, Jiyun; Qiu, Qiaoyun; Wang, Shuwei; Wong, Kam-Fai; Cheung, R.

doi:10.1007/978-3-642-41644-6_10

Cited by 12 publications

(11 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Note that in Table 3, the top performer of NLP&CC 2013 CLOA evaluation is the HLT-HITSZ system(underscored in the table), which used the co-training method in transfer learning (Gui et al, 2013), proving that co-training is quite effective for cross-lingual analysis. With the additional negative transfer detection, our proposed approach achieves the best performance on this dataset outperformed the top system (by HLT-HITSZ) by a 2.97% which translate to 13.1% error reduction improvement to this state-of-the-art system as shown in the last row of To further investigate the effectiveness of our method, the third set of experiments evaluate the negative transfer detection (NTD) compared to co-training (CO) without negative transfer detection as shown in Table.4 and Fig.3 Taking all categories of data, our proposed method improves the overall average precision (the best cases) from 79.4% to 80.1% when compared to the state of the art system which translates to error reduction of 3.40% (pvalue≤0.01 in Wilcoxon signed rank test).…”

Section: Cloa Experiments Resultsmentioning

confidence: 99%

Cross-lingual Opinion Analysis via Negative Transfer Detection

Gui¹,

Xu²,

Lu³

et al. 2014

Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)

Self Cite

View full text Add to dashboard Cite

Transfer learning has been used in opinion analysis to make use of available language resources for other resource scarce languages. However, the cumulative class noise in transfer learning adversely affects performance when more training data is used. In this paper, we propose a novel method in transductive transfer learning to identify noises through the detection of negative transfers. Evaluation on NLP&CC 2013 cross-lingual opinion analysis dataset shows that our approach outperforms the state-of-the-art systems. More significantly, our system shows a monotonic increase trend in performance improvement when more training data are used.

show abstract

Section: Cloa Experiments Resultsmentioning

confidence: 99%

Cross-lingual Opinion Analysis via Negative Transfer Detection

Gui¹,

Xu²,

Lu³

et al. 2014

Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)

Self Cite

View full text Add to dashboard Cite

show abstract

“…To deal with this issue, several methods have been proposed to reduce translation errors, such as applying both directions of translation simultaneously (Hajmohammadi et al, 2014) or enriching the MT system with sentiment patterns (Hiroshi et al, 2004). In the case of supervised systems, self-training and co-training techniques have also been explored to improve performance (Gui et al, 2013;Gui et al, 2014).…”

Section: Multilingual Samentioning

confidence: 99%

Universal, unsupervised (rule-based), uncovered sentiment analysis

Vilares

Gómez-Rodríguez

Alonso

2017

Knowledge-Based Systems

View full text Add to dashboard Cite

We present a novel unsupervised approach for multilingual sentiment analysis driven by compositional syntax-based rules. On the one hand, we exploit some of the main advantages of unsupervised algorithms: (1) the interpretability of their output, in contrast with most supervised models, which behave as a black box and (2) their robustness across different corpora and domains. On the other hand, by introducing the concept of compositional operations and exploiting syntactic information in the form of universal dependencies, we tackle one of their main drawbacks: their rigidity on data that are structured differently depending on the language concerned. Experiments show an improvement both over existing unsupervised methods, and over state-of-the-art supervised models when evaluating outside their corpus of origin. Experiments also show how the same compositional operations can be shared across languages. The system is available at http://www.grupolys. org/software/UUUSA/.

show abstract

“…Then, they performed training and testing in two independent views: English view and Chinese view. Gui et al (2013) combined self-training approach with co-training approach by estimating the confidence of each monolingual system. Li et al (2013) selected the samples in the source language that were similar to those in the target language to decrease the gap between two languages.…”

Section: Cross-language Sentiment Classification (Clsc)mentioning

confidence: 99%

“…Gui et al (2013) combined self-training approach with co-training approach by estimating the confidence of each monolingual system. Li et al (2013) selected the samples in the source language that were similar to those in the target language to decrease the gap between two languages. Zhou et al (2014a) proposed a combination CLSC model, which adopted denoising autoencoders (Vincent et al, 2008) to enhance the robustness to translation errors of the input.…”

Section: Cross-language Sentiment Classification (Clsc)mentioning

confidence: 99%

“…To improve the classification accuracy, multiview approaches have been proposed. In these approaches, the resources in the source language and their translations in the target language are both used to train sentiment classifiers in two independent views (Wan, 2009;Gui et al, 2013;Zhou et al, 2014a). The final results are determined by ensemble classifiers in these two views to overcome the weakness of monolingual classifiers.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning Bilingual Sentiment Word Embeddings for Cross-language Sentiment Classification

Zhou¹,

Chen²,

Shi³

et al. 2015

Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Confere

View full text Add to dashboard Cite

The sentiment classification performance relies on high-quality sentiment resources. However, these resources are imbalanced in different languages. Cross-language sentiment classification (CLSC) can leverage the rich resources in one language (source language) for sentiment classification in a resource-scarce language (target language). Bilingual embeddings could eliminate the semantic gap between two languages for CLSC, but ignore the sentiment information of text. This paper proposes an approach to learning bilingual sentiment word embeddings (BSWE) for English-Chinese CLSC. The proposed B-SWE incorporate sentiment information of text into bilingual embeddings. Furthermore, we can learn high-quality BSWE by simply employing labeled corpora and their translations, without relying on largescale parallel corpora. Experiments on NLP&CC 2013 CLSC dataset show that our approach outperforms the state-of-theart systems.

show abstract

A Mixed Model for Cross Lingual Opinion Analysis

Cited by 12 publications

References 5 publications

Cross-lingual Opinion Analysis via Negative Transfer Detection

Cross-lingual Opinion Analysis via Negative Transfer Detection

Universal, unsupervised (rule-based), uncovered sentiment analysis

Learning Bilingual Sentiment Word Embeddings for Cross-language Sentiment Classification

Contact Info

Product

Resources

About