LPIS Home Page
Google Search

Title: Correlation-Based Pruning of Stacked Binary Relevance Models for Multi-Label Learning
Author(s): G. Tsoumakas, A. Dimou, E. Spyromitros-Xioufis, V. Mezaris, I. Kompatsiaris, I. Vlahavas.
Availability: Click here to download the PDF (Acrobat Reader) file (16 pages).
Appeared in: Proceedings of the 1st International Workshop on Learning from Multi-Label Data (MLD'09), G. Tsoumakas, Min-Ling Zhang, Zhi-Hua Zhou (Ed.), pp. 101-116, Bled, Slovenia, 2009.
Abstract: Binary relevance (BR) learns a single binary model for each different label of multi-label data. It has linear complexity with respect to the number of labels, but does not take into account label correlations and may fail to accurately predict label combinations and rank labels according to relevance with a new instance. Stacking the models of BR in order to learn a model that associates their output to the true value of each label is a way to alleviate this problem. In this paper we propose the pruning of the models participating in the stacking process, by explicitly measuring the degree of label correlation using the phi coefficient. Exploratory analysis of phi shows that the correlations detected are meaningful and useful. Empirical evaluation of the pruning approach shows that it leads to substantial reduction of the computational cost of stacking and occasional improvements in predictive performance.
See also :

        This paper has been cited by the following:

1 Wang, Z., Hu, Y., and Chia, L. 2010. Multi-label learning by Image-to-Class distance for scene classification and image annotation. In Proceedings of the ACM international Conference on Image and Video Retrieval (Xi'an, China, July 05 - 07, 2010). CIVR '10. ACM, New York, NY, 105-112.
2 Zheng, W., Wang, C.-K., Liu, Z., Wang, J.-M. (2010) A multi-label classification algorithm based on random walk model, Jisuanji Xuebao/Chinese Journal of Computers, 33(8), August 2010, Pages 1418-1426
3 Read, J. (2010) Scalable Multi-Label Classification, PhD Thesis, University of Waikato.
4 Tenenboim-Chekina, L., Rokach, L., Shapira, B. (2010) Identification of Label Dependencies for Multi-label Classification, Proc. 2nd International Workshop on Multi-Label Learning.
5 Chekina, L., Rokach, L., Shapira, B. (2011) Meta-learning for selecting a multi-label classification algorithm, Proceedings - IEEE International Conference on Data Mining, ICDM, art. no. 6137383, pp. 220-227.
6 Wang, Hua; Huang, Heng; Ding, Chris; (2011) Image annotation using bi-relational graph of images and semantic labels, Proc. 2011 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.793-800, 20-25 June 2011
7 Montañés, E., Quevedo, J.R., del Coz, J.J. (2011) Aggregating Independent and Dependent Models to Learn Multi-label Classifiers, Proc. ECML PKDD 2011, part II, pp. 484-500.
8 Bielza, C., Li, G., Larrañaga, P. Multi-dimensional classification with Bayesian networks (2011) International Journal of Approximate Reasoning, 52 (6), pp. 705-727.
9 Sheng-Jun Huang, Zhi-Hua Zhou. Multi-label learning by exploiting label correlations locally. In: Proceedings of the 26th AAAI Conference on Artificial Intelligence (AAAI'12), Toronto, Canada, 2012.
10 Sheng-Jun Huang, Yang Yu, and Zhi-Hua Zhou. 2012. Multi-label hypothesis reuse. In Proceedings of the 18th ACM SIGKDD international conference on Knowledge discovery and data mining (KDD '12). ACM, New York, NY, USA, 525-533.
11 Wicker, J., Pfahringer, B., Kramer, S. (2012) Multi-label classification using boolean matrix decomposition, Proceedings of the ACM Symposium on Applied Computing, pp. 179-186.
12 Fu, B., Wang, Z., Pan, R., Xu, G., Dolog, P. (2012) Learning tree structure of label dependency for multi-label learning, Proc 16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012, 7301 LNAI (PART 1), pp. 159-170.