Abstract: Automatic classification of sentiment data (e.g., reviews, blogs) has many applications in enterprise user management systems, and can help us understand people&x2019;s attitudes about products or services. However, it is difficult to train an accurate sentiment classifier for different domains. One of the major reasons is that people often use different words to express the same sentiment in different domains, and we cannot easily find a direct mapping relationship between them to reduce the differences between domains. So, the accuracy of the sentiment classifier will decline sharply when we apply a classifier trained in one domain to other domains. In this paper, we propose a novel approach called words alignment based on association rules (WAAR) for cross-domain sentiment classification, which can establish an indirect mapping relationship between domain-specific words in different domains by learning the strong association rules between domain-shared words and domain-specific words in the same domain. In this way, the differences between the source domain and target domain can be reduced to some extent, and a more accurate cross-domain classifier can be trained. Experimental results on Amazon^® datasets show the effectiveness of our approach on improving the performance of cross-domain sentiment classification.

基于关联规则进行词对齐的跨领域情感分类算法

[1]Agrawal R, Srikant R, 1994. Fast algorithms for mining association rules. 20th Int Conf on Very Large Data Bases, 15(6):487-499.

[2]Ando R, Zhang T, 2005. A framework for learning predictive structures from multiple tasks and unlabeled data. J Mach Learn Res, 6:1817-1853.

[3]Balazs J, Velasquez J, 2016. Opinion mining and information fusion: a survey. Inf Fusions, 27(C):95-110.

[4]Blitzer J, Dredze M, Pereira F, 2007. Biographies, bollywood, boom-boxes and blenders: domain adaptation for sentiment classification. 45th Annual Meeting of the Association of Computational Linguistics, p.440-447.

[5]Bollegala D, Weir D, Carroll J, 2013. Cross-domain sentiment classification using a sentiment sensitive thesaurus. IEEE Trans Know Data Eng, 25(8):1719-1731.

[6]Chen M, Xu Z, Kilian Q, et al., 2012. Marginalized denoising autoencoders for domain adaptation. arXiv preprint, cs.LG(1206.4683):1-8.

[7]Chung FRK, 1997. Spectral Graph Theory. Co-publication of the AMS and CBMS.

[8]Dave K, Lawrence S, Pennock D, 2003. Mining the peanut gallery: opinion extraction and semantic classification of product reviews. 12th Int Conf on World Wide Web, p.519-528.

[9]Ding S, Jia H, Shi Z, 2014. Spectral clustering algorithm based on adaptive Nystrom sampling for big data analysis. J Softw, 25(9):2037-2049.

[10]Glorot X, Bordes A, Bengio Y, 2011. Domain adaptation for large-scale sentiment classification: a deep learning approach. 28th Int Conf on Machine Learning, p.513-520.

[11]Goldberg A, Zhu X, 2006. Seeing stars when there are not many stars: graph-based semi-supervised learning for sentiment categorization. 1st Workshop on Graph Based Methods for Natural Language Processing, p.45-52.

[12]Jiang J, Zhai C, 2007. Instance weighting for domain adaptation in NLP. 45th Annual Meeting of the Association of Computational Linguistics, p.264-271.

[13]Li L, Jin X, Long M, 2012. Topic correlation analysis for cross-domain text classification. AAAI Conf on Artificial Intelligence, p.998-1004.

[14]Li T, Zhang Y, Sindhwani V, 2009. A non-negative matrix tri-factorization approach to sentiment classification with lexical prior knowledge. Joint Conf of the 47th Annual Meeting of the ACL and the 4th Int Joint Conf on Natural Language Processing, p.244-252.

[15]Pan S, Ni X, Sun J, et al., 2010. Cross-domain sentiment classification via spectral feature alignment. 19th Int Conf on World Wide Web, p.751-760.

[16]Pang B, Lee L, Vaithyanathan S, 2002. Thumbs up? sentiment classification using machine learning techniques. ACL-02 Conf on Empirical Methods in Natural Language Processing, p.79-86.

[17]Pantelis A, Loannis K, Loannis K, et al., 2016. Learning patterns for discovering domain oriented opinion words. Know Inf Syst, 2017(1):1-33.

[18]Schölkopf B, J Platt TH, 2007. Correcting sample selection bias by unlabeled data. Advances in Neural Information Processing Systems, p.601-608.

[19]Turney PD, 2002. Thumbs up or thumbs down? semantic orientation applied to unsupervised classification of reviews. 40th Annual Meeting on Association for Computational Linguistics, p.417-424.

[20]Wu Q, Tan S, Xu H, et al., 2010a. Cross-domain opinion analysis based on random-walk model. J Comput Res Dev, 47(12):2123-2131.

[21]Wu Q, Tan S, Zhang G, et al., 2010b. Research on cross-domain opinion analysis. J Chin Inf Process, 24(1):77-83.

[22]Yang Y, Pedersen J, 1997. A comparative study on feature selection in text categorization. 14th Int Conf on Machine Learning, p.412-420.

[23]Zadrozny B, 2004. Learning and evaluating classifiers under sample selection bias. 21st Int Conf on Machine Learning, p.114-121.

[24]Zhou G, Zhou Y, Guo X, et al., 2015. Cross-domain sentiment classification via topical correspondence transfer. Neurocomputing, 159:298-305.

[25]Zhuang L, Jing F, Zhu X, 2006. Movie review mining and summarization. 15th ACM Int Conf on Information and Knowledge Management, p.43-50.