skent259 / omisvm-sims

Simulation code for "Ordinal Multiple Instance Support Vector Machines"

Simulations for Ordinal Multiple Instance Support Vector Machines

Striving for reproducibility, this repo details the simulations for Ordinal Multiple Instance Support Vector Machines by Sean Kent and Menggang Yu. This allows for recreation of all figures and tables used in the paper, except in the case of proprietary data. Full details and descriptions are available in the paper.

As a quick reference for getting started:

• The experiments can be run via the run.sh file, however this will take considerable CPU time. We recommend that you run in a high-throughput environment in batches (see run.sh for inspiration)
• Figures from the paper can be recreated by running analysis/create-all-figures.R.
• Data (open-source) can be downloaded by running data/build-data.sh.
• Simulation source code can be found in the sim/ directory.
• The condor/ directory can be ignored. It was used to run simulations in the HTCondor (high-throughput) environment.

Methods Compared

The main model code relies on the mildsvm package (v0.4.0) in R, which can be found here: https://github.com/skent259/mildsvm

Method name	Function	parameters	Reference
OMI-SVM (proposed)	mildsvm::omisvm	method = "qp-heuristic"	[1]
MI-SVM (OVA)	mildsvm::misvm_orova	method = "qp-heuristic"	[2], [3]
SI-SVOREXC	mildsvm::svor_exc	method = "smo"	[4], [5]
MIOR (as written)	mildsvm::mior	method = "qp-heuristic", option = "xiao"	[6]
MIOR (corrected)	mildsvm::mior	method = "qp-heuristic", option = "corrected"	[6]

Experiments

Data set	Experiment type	Name	ID	Section
Amazon review ratings	Performance	per-amrev	1.0	Evaluation
tissue micro-array	Performance	per-tma	2.0	Evaluation
IMDB movie reviews	Training data size	size-imdb	3.0	Evaluation
SWD	Training data size	size-swd	4.0	Appendix
winequality-red	Training data size	size-wq	5.0	Appendix
CAR	Witness rate changes	wr-car	6.0	Appendix
ERA	Witness rate changes	wr-era	7.0	Appendix

Most data from the experiments is available from open-source repositories:

• Amazon review ratings: http://sifaka.cs.uiuc.edu/~wang296/Data/ (Wang et al. 2011)
• IMDB movie reviews: https://ai.stanford.edu/~amaas/data/sentiment/ (Maas et al. 2011)
• winequality-red: https://archive.ics.uci.edu/ml/datasets/Wine+Quality (Cortez et al. 2009)
• SWD: https://www.openml.org/d/1028
• CAR: https://www.openml.org/d/40975 (Bohanec and Rajkovič, 1988; Dua and Graff, 2017)
• ERA: https://www.openml.org/d/1030

The TMA data set is not publicly available due to privacy restrictions. It can be made available from the authors.

References

[1] Kent, S., & Yu, M. (2023+). Ordinal multiple instance support vector machines. In prep.

[2] Andrews, S., Tsochantaridis, I., & Hofmann, T. (2003). Support vector machines for multiple-instance learning. Advances in Neural Information Processing Systems 15, 577–584).

[4] Hsu, C.-W., & Lin, C.-J. (2002). A comparison of methods for multiclass support vector machines. IEEE Transactions on Neural Networks, 13(2), 415–425. https://doi.org/10.1109/72.991427

[4] Chu, W., & Keerthi, S. S. (2007). Support vector ordinal regression. Neural Computation, 19(3), 792–815. https://doi.org/10.1162/neco.2007.19.3.792

[5] Alpaydın, E., Cheplygina, V., Loog, M., & Tax, D. M. J. (2015). Single- vs. Multiple-instance classification. Pattern Recognition, 48(9), 2831–2838. https://doi.org/10.1016/j.patcog.2015.04.006

[6] Xiao, Y., Liu, B., & Hao, Z. (2018). Multiple-instance ordinal regression. IEEE Transactions on Neural Networks and Learning Systems, 29(9), 4398–4413. https://doi.org/10.1109/TNNLS.2017.2766164

[7] Wang, H., Wang, C., Zhai, C., & Han, J. (2011). Learning online discussion structures by conditional random fields. Proceedings of the 34th International Acm Sigir Conference on Research and Development in Information Retrieval, 435–444. https://doi.org/10.1145/2009916.2009976

[8] Maas, A., Daly, R. E., Pham, P. T., Huang, D., Ng, A. Y., & Potts, C. (2011). Learning word vectors for sentiment analysis. Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, 142–150.

[9] Cortez, P., Cerdeira, A., Almeida, F., Matos, T., & Reis, J. (2009). Modeling wine preferences by data mining from physicochemical properties. Decision Support Systems, 47(4), 547–553. https://doi.org/10.1016/j.dss.2009.05.016

[10] Bohanec, M., & Rajkovič, V. (1988). Knowledge acquisition and explanation for multi-attribute decision. 8th International Workshop on Expert Systems and Their Applications, 1, 59–78.

[11] Dua, D., & Graff, C. (2017). UCI Machine Learning Repository. University of California, Irvine, School of Information and Computer Sciences. http://archive.ics.uci.edu/ml