Online Semi-supervised Learning for Multi-target Regression in Data Streams Using AMRules
Most data streams systems that use online Multi-target regression yield vast amounts of data which is not targeted. Targeting this data is usually impossible, time consuming and expensive. Semi-supervised algorithms have been proposed to use this untargeted data (input information only) for model improvement. However, most algorithms are adapted to work on batch mode for classification and require huge computational and memory resources.
Therefore, this paper proposes an semi-supervised algorithm for online processing systems based on AMRules algorithm that handle both targeted and untargeted data and improves the regression model. The proposed method was evaluated through a comparison between a scenario where the untargeted examples are not used on the training and a scenario where some untargeted examples are used. Evaluation results indicate that the use of the untargeted examples improved the target predictions by improving the model.
KeywordsMulti-target regression Semi-supervised learning AMRules Data streams
This work was partly supported by the European Commission through MAESTRA (ICT-2013-612944) and the Project TEC4Growth - Pervasive Intelligence, Enhancers and Proofs of Concept with Industrial Impact/NORTE-01-0145-FEDER-000020 is financed by the North Portugal Regional Operational Programme (NORTE 2020), under the PORTUGAL 2020 Partnership Agreement, and through the European Regional Development Fund (ERDF).
- 2.Levatic, J., Ceci, M., Kocev, D., Dzeroski, S.: Semi-supervised learning for multi-target regression. In: Third International Workshop, NFMCP, Held in Conjunction with ECML-PKDD, pp. 3–18 (2014)Google Scholar
- 4.Duarte J., Gama, J.: Multi-target regression from high-speed data streams with adaptive model rules. In: IEEE Conference on Data Science and Advanced Analytics (2015)Google Scholar
- 5.Goldberg, A.B., Zhu, X., Furger, A., Jun-Ming, X.: OASIS: online active semi-supervised learning. In: Proceedings of the Twenty-Fifth AAAI Conference on Artificial Intelligence, AAAI, San Francisco, California, USA, 7–11 August 2011Google Scholar
- 7.Ozoh, P., Abd-rahman, S., Labadin, J., Apperley, M.: Article: a comparative analysis of techniques for forecasting electricity consumption. Int. J. Comput. Appl. 88(15), 8–12 (2014)Google Scholar
- 8.Chalabi, Z., Mangtani, P., Hashizume, M., Imai, C., Armstrong, B.: Article: time series regression model for infectious disease and weather. Int. J. Environ. Res. 142, 319–327 (2015)Google Scholar
- 10.Ariyo, A.A., Adewumi, A.O., Ayo, C.K.: Stock price prediction using the arima model. In: Proceedings of the UKSim-AMSS 16th International Conference on Computer Modelling and Simulation, UKSIM 2014, pp. 106–112, Washington, DC, USA. IEEE Computer Society (2014)Google Scholar
- 11.Chapelle, O., Schlkopf, B., Zien, A.: Semi-Supervised Learning, 1st edn. The MIT Press, Cambridge (2010)Google Scholar
- 12.Albalate, A., Minker, W.: Semi-supervised and Unsupervised Machine Learning. ISTE/Wiley, London (2011)Google Scholar
- 14.Radosavljevic, V., Vucetic, S., Obradovic, Z.: Continuous conditionalrandom fields for regression in remote sensing. In: 19th European Conference on Artificial Intelligence, Proceedings of the 2010 Conference on ECAI 2010, pp. 809–814, Amsterdam, The Netherlands. IOS Press (2010)Google Scholar
- 15.Stojanovic, J., Jovanovic, M., Gligorijevic, D., Obradovic, Z.: Semi-supervised learning for structured regression on partially observed attributed graphs. In: SIAM International Conference on Data Mining (SDM) (2015)Google Scholar
- 18.Chen, W.: Passive, Active, and Digital Filters, 3rd edn. CRC Press, Baco Raton (2009)Google Scholar