Abstract
\(R^2\), despite being a widely used goodness-of-fit measure for linear regression shows erratic behavior in presence of data contamination. Several alternate measures have been proposed that show some improvement under specific conditions. However, no single universal measure exists as such that can be used to assess and compare performance of linear regression models without being concerned about composition of data. This paper proposes a new robust \(R^2\) measure that is found to work better than existing measures across scenarios. Performance superiority has been demonstrated using extensive simulation results and three real publicly available datasets. Proposed methodology also shows significant improvement in outlier detection and comparable performance to other established methods for robust linear regression.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Peter, J.R.: Least median of squares regression. J. Am. Stat. Assoc. 79(388), 871–880 (1984)
James, P.S.: Outliers and influential data points in regression analysis. Psychol. Bull. 95(2), 334–344 (1984)
John, M.S., William, L.S.: Algorithms and complexity for least median of squares regression. Discrete Appl. Math. 14(1), 93–100 (1986)
Santiago, V.: On the behaviour of residual plots in robust regression. Stat. Econometrics Ser. 04, 93–104 (1993). Working Paper
I-Cheng, Y.: Modeling of strength of high performance concrete using artificial neural networks. Cem. Concr. Res. 28(12), 1797–1808 (1998)
Victoria, J.H., Jim, A.: A survey of outlier detection methodologies. Artif. Intell. Rev. 22(2), 85–126 (2004)
Peter, J.R., Annick, M.L.: Robust Regression and Outlier Detection, vol. 589. Wiley, New York (2005)
Jeff, T.T., Joseph, W.M.: Rank-based analysis of linear models using R. J. Stat. Softw. 14(7), 1–26 (2005)
Heysem, K., Pınar, T., Fikret, S.G.: Local and global learning methods for predicting power of a combined gas and steam turbine. In: Proceedings of the International Conference on Emerging Trends in Computer and Electronics Engineering (ICETCEE 2012), Dubai, UAE, pp. 13–18 (2012)
Khaleelur Rahman, S.M.A., Mohamed Sathik, M., Senthamarai Kannan, K.: Multiple linear regression models in outlier detection. Int. J. Res. Comput. Sci. 2(2), 23–28 (2012)
Moshe, L.: UCI Machine Learning Repository. Irvine, CA: University of California, School of Information and Computer Science (2013). http://archive.ics.uci.edu/ml
Pınar, T.: Prediction of full load electrical power output of a base load operated combined cycle power plant using machine learning methods. Int. J. Electr. Power Energy Syst. 60, 126–140 (2014). ISSN 0142–0615
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this paper
Cite this paper
Deb, S. (2017). A Novel Robust R-Squared Measure and Its Applications in Linear Regression. In: Phon-Amnuaisuk, S., Au, TW., Omar, S. (eds) Computational Intelligence in Information Systems. CIIS 2016. Advances in Intelligent Systems and Computing, vol 532. Springer, Cham. https://doi.org/10.1007/978-3-319-48517-1_12
Download citation
DOI: https://doi.org/10.1007/978-3-319-48517-1_12
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-48516-4
Online ISBN: 978-3-319-48517-1
eBook Packages: EngineeringEngineering (R0)