Abstract
A data classification model in which the average mutual information between source objects and made decisions is a function of the error probability is investigated. Optimization of the model consists in finding a tradeoff “mutual information–error probability” (MIEP) relation between the minimal average mutual information and the error probability, which is analogous to the well-known rate distortion function for source coding with a given fidelity in the case of a noisy observation channel. A lower bound for the MIEP relation is constructed, which provides a lower bound for the classification error probability on a given set of objects for any fixed value of the average mutual information. The MIEP relation and its lower bound are generalized to ensembles of sources. The obtained bounds are useful for estimating the error probability redundancy for decision algorithms with given sets of discriminant functions.
Similar content being viewed by others
REFERENCES
R. G. Gallager, Information Theory and Reliable Communication (Wiley, New York, 1968).
L. I. Kuncheva, C. J. Whitaker, C. A. Shipp, and R. P. W. Duin, “Limits on the majority vote accuracy in classifier fusion,” Pattern Anal. Appl. 6, 22–31 (2003). https://doi.org/10.1007/s10044-002-0173-7
L. Lam and C. Y. Suen, “Application of majority voting to pattern recognition: An analysis of its behavior and performance,” IEEE Trans. Syst. Man Cybern. 27 (5), 553–568 (1997). https://doi.org/10.1109/3468.618255
D. J. C. MacKay, Information Theory, Inference, and Learning Algorithms (Cambridge Univ. Press, Cambridge, 2003).
R. L. Dobrushin and B. S. Tsybakov, “Information transmission with additional noise,” IRE Trans. Inf. Theory 8 (5), 293–304 (1962). https://doi.org/10.1109/TIT.1962.1057738
M. Lange, S. Ganebnykh, and A. Lange, “An information approach to accuracy comparison for classification schemes in an ensemble of data sources,” Intelligent Data Processing (Springer, Berlin, 2019), Vol. 794, pp. 28–43. https://doi.org/10.1007/978-3-030-35400-8_3.
M. M. Lange, S. N. Ganebnykh, and A. M. Lange, “On an information-theoretical lower bound to a classification error probability,” Abstracts of the 19th Russian Conference on Mathematical Methods for Pattern Recognition (MMPR-19) (Moscow, 2019).
R. O. Duda, P. E. Hart, and D. G. Stork, Pattern Classification, 2nd ed. (Wiley, New York, 2001).
M. M. Lange and S. N. Ganebnykh, “On fusion schemes for multiclass object classification with reject in a given ensemble of sources,” J. Phys. Conf. Ser. 1096 (012048) (2018). https://doi.org/10.1088/1742-6596/1096/1/012048
"Distance matrices for sign images." http://sourceforge.net/projects/distance-matrices-signature. Accessed June 2020.
"Distance matrices for face images." http://sourceforge.net/projects/distance-matrices-face. Accessed June 2020.
Funding
This work was supported in part by the Russian Foundation for Basic Research, project no. 18-07-01231.
Author information
Authors and Affiliations
Corresponding authors
Additional information
Translated by I. Ruzanova
Rights and permissions
About this article
Cite this article
Lange, A.M., Lange, M.M. & Paramonov, S.V. Tradeoff Relation between Mutual Information and Error Probability in Data Classification Problem. Comput. Math. and Math. Phys. 61, 1181–1193 (2021). https://doi.org/10.1134/S0965542521070113
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1134/S0965542521070113