Abstract: Selecting pivot features that connect a source domain to a target domain is an important first step in unsupervised domain adaptation (UDA). Although different strategies such as the frequency of a feature in a domain, mutual (or pointwise mutual) information have been proposed in prior work in domain adaptation (DA) for selecting pivots, a comparative study into (a) how the pivots selected using existing strategies differ, and (b) how the pivot selection strategy affects the performance of a target DA task remain unknown. In this paper, we perform a comparative study covering different strategies that use both labelled (available for the source domain only) as well as unlabelled (available for both the source and target domains) data for selecting pivots for UDA. Our experiments show that in most cases pivot selection strategies that use labelled data outperform their unlabelled counterparts, emphasising the importance of the source domain labelled data for UDA. Moreover, pointwise mutual information and frequency-based pivot selection strategies obtain the best performances in two state-of-the-art UDA methods.
Blitzer J., Dredze M. & Pereira F. 2007. Biographies, bollywood, boom-boxes and blenders: Domain adaptation for sentiment classification. In Proceedings of the ACL, 440–447.
Bollegala D., Mu T. & Goulermas J. Y. 2015. Cross-domain sentiment classification using sentiment sensitive embeddings. IEEE Transactions on Knowledge and Data Engineering 28(2), 398–410.
Bollegala D., Weir D. & Carroll J. 2011. Using multiple sources to construct a sentiment sensitive thesaurus for cross-domain sentiment classification. In Proceedings of the ACL, 132–141.
Koehn P. & Schroeder J. 2007. Experiments in domain adaptation for statistical machine translation. In Proceedings of the Second Workshop on Statistical Machine Translation, 224–227.
Mansour R. H., Refaei N., Gamon M., Sami K. & Abdel-Hamid A. 2013. Revisiting the old kitchen sink: do we need sentiment domain adaptation? In Proceedings of the RANLP, 420–427.
Pan S. J., Ni X., Sun J.-T., Yang Q. & Chen Z. 2010. Cross-domain sentiment classification via spectral feature alignment. In Proceedings of WWW, 751–760.
Yu J. & Jiang J. 2015. A hassle-free unsupervised domain adaptation method using instance similarity features. In Proceedings of the ACL-IJCNLP, 168–173.
Zhang Y., Xu X. & Hu X. 2015. A common subspace construction method in cross-domain sentiment classification. In Procedings of International Conference on Electronic Science and Automation Control (ESAC), 48–52.
Xia Cui, Noor Al-Bazzaz, Danushka Bollegala, Frans Coenen. 2018. A comparative study of pivot selection strategies for unsupervised cross-domain sentiment classification. The Knowledge Engineering Review 33(1), doi: 10.1017/S0269888918000085
Xia Cui, Noor Al-Bazzaz, Danushka Bollegala, Frans Coenen. 2018. A comparative study of pivot selection strategies for unsupervised cross-domain sentiment classification. The Knowledge Engineering Review 33(1), doi: 10.1017/S0269888918000085
Abstract: Abstract: Selecting pivot features that connect a source domain to a target domain is an important first step in unsupervised domain adaptation (UDA). Although different strategies such as the frequency of a feature in a domain, mutual (or pointwise mutual) information have been proposed in prior work in domain adaptation (DA) for selecting pivots, a comparative study into (a) how the pivots selected using existing strategies differ, and (b) how the pivot selection strategy affects the performance of a target DA task remain unknown. In this paper, we perform a comparative study covering different strategies that use both labelled (available for the source domain only) as well as unlabelled (available for both the source and target domains) data for selecting pivots for UDA. Our experiments show that in most cases pivot selection strategies that use labelled data outperform their unlabelled counterparts, emphasising the importance of the source domain labelled data for UDA. Moreover, pointwise mutual information and frequency-based pivot selection strategies obtain the best performances in two state-of-the-art UDA methods.
HTML
Acknowledgments
The authors would like to thank all the anonymous reviewers, and the support from editors.
Note that the original proposal by Blitzer et al. (2007) was to use mutual information with source domain labelled data as we discuss later in Section 3.2. However, for comparison purposes we define a pivothood score based on frequency and source domain labelled data here.
Xia Cui, Noor Al-Bazzaz, Danushka Bollegala, Frans Coenen. 2018. A comparative study of pivot selection strategies for unsupervised cross-domain sentiment classification. The Knowledge Engineering Review 33(1), doi: 10.1017/S0269888918000085
Xia Cui, Noor Al-Bazzaz, Danushka Bollegala, Frans Coenen. 2018. A comparative study of pivot selection strategies for unsupervised cross-domain sentiment classification. The Knowledge Engineering Review 33(1), doi: 10.1017/S0269888918000085
Catalog
Share:
Export File
Citation
Xia Cui, Noor Al-Bazzaz, Danushka Bollegala, Frans Coenen. 2018. A comparative study of pivot selection strategies for unsupervised cross-domain sentiment classification. The Knowledge Engineering Review 33(1), doi: 10.1017/S0269888918000085
Xia Cui, Noor Al-Bazzaz, Danushka Bollegala, Frans Coenen. 2018. A comparative study of pivot selection strategies for unsupervised cross-domain sentiment classification. The Knowledge Engineering Review 33(1), doi: 10.1017/S0269888918000085
