Abstract
We consider a scenario where an automatic classifier has been built, but it sometimes decides to ask the correct label of an instance to an oracle, instead of accepting its own prediction. This interactive classifier only knows with certainty the labels provided by the oracle. Our proposal is to use this information to dynamically improve the behavior of the classifier, either increasing its accuracy when it is being used autonomously or reducing the number of queries to the oracle. We have tested our proposal by using twenty data sets and two adaptive classifiers from the Massive Online Analysis (MOA) open source framework for data stream mining.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
The evaluation should not be based only on its predictive accuracy but should also take into account the cost of the human intervention.
- 2.
In [1], it was shown that with values of \(\beta \) lesser than or equal to 0.5, the interactive classifiers systematically outperform their non-interactive counterparts.
- 3.
In this paper we assume that the process is stationary and no concept drift occurs. The case of non-stationary problems, where a concept drift can modify the probability distribution, will be considered in future research.
- 4.
By lack of space, we omit the proof of this result.
- 5.
The new instances are supposed to come from the same data distribution.
- 6.
Without taking into account those instances labeled by the oracle, \(n_i\).
References
Acid, S., de Campos, L.M., Fernández, M.: Evaluation methods and strategies for the interactive use of classifiers. Int. J. Hum.-Comput. Stud. 70(5), 321–331 (2012)
Agrawal, R., Imielinski, T., Swami, A.: Database mining: a performance perspective. IEEE Trans. Knowl. Data Eng. 5(6), 914–925 (1993)
Auer, P., Cesa-Bianchi, N., Gentile, C.: Adaptive and self-confident on-line learning algorithms. J. Comput. Syst. Sci. 64, 48–75 (2002)
Barlet, P.L., Wegkamp, M.H.: Classification with a reject option using a hinge loss. J. Mach. Learn. Res. 9, 1823–1840 (2008)
Bifet, A., Holmes, G., Kirkby, R., Pfahringer, B.: MOA: massive online analysis. J. Mach. Learn. Res. 11, 1601–1604 (2010)
Fu, Y., Zhu, X., Li, B.: A survey on instance selection for active lerning. Knowl. Inf. Syst. 35, 249–283 (2013)
Fumera, G., Roli, F., Giacinto, G.: Reject option with multiple thresholds. Pattern Recogn. 33(12), 2099–2101 (2000)
Gama, J.: Knowledge Discovery from Data Streams. Chapman and Hall/CRC, Boca Raton (2010)
Gama, J.: A survey on learning from data streams: current and future trends. Prog. Artif. Intell. 1, 45–55 (2012)
Hulten, G., Spencer, L., Domingos, P.: Mining time-changing data streams. In: KDD-01, pp. 97–106. ACM Press (2001)
Lewis, D., Gale, W.: A sequential algorithm for training text classifiers. In: Proceedings of the ACM SIGIR Conference, pp. 3–12 (1994)
Settles, B.: Active learning literature survey. Computer Sciences Technical Report 1648, University of Wisconsin-Madison (2009)
Shalev-Shwartz, S.: Online learning and online convex optimization. Found. Trends Mach. Learn. 4(2), 107–194 (2012)
van Rijsbergen, C.J.: Foundation of evaluation. J. Documentation 30(4), 365–373 (1974)
Acknowledgements
This work has been funded by the Spanish Ministry of Economy and Competitiveness under the project TIN2013-42741-P and the European Regional Development Fund (ERDF-FEDER).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this paper
Cite this paper
Acid, S., de Campos, L.M. (2016). Improving Automatic Classifiers Through Interaction. In: Rutkowski, L., Korytkowski, M., Scherer, R., Tadeusiewicz, R., Zadeh, L., Zurada, J. (eds) Artificial Intelligence and Soft Computing. ICAISC 2016. Lecture Notes in Computer Science(), vol 9693. Springer, Cham. https://doi.org/10.1007/978-3-319-39384-1_1
Download citation
DOI: https://doi.org/10.1007/978-3-319-39384-1_1
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-39383-4
Online ISBN: 978-3-319-39384-1
eBook Packages: Computer ScienceComputer Science (R0)