Informatics and Applications
2017, Volume 11, Issue 2, pp 74-84
ON MODIFICATION OF THE MEAN SQUARED ERROR LOSS FUNCTION FOR SOLVING NONLINEAR HETEROSCEDASTIC ERRORS-IN-VARIABLES PROBLEMS
Abstract
The paper considers the problem of finding the optimal parameters of a nonlinear regression model accounting for errors in both dependent and independent variables. The errors of different measurements are assumed to belong to different probability distributions with different variances. A modified mean squared error- based loss function is derived and analyzed for this case. In the computational experiment, the measurements of the laser's radiation power as a nonlinear function of the resonator's transparency are used to compare the parameters vectors minimizing the presented loss function and the classical mean squared error. The convergence of the parameters minimizing the presented loss function to the optimal parameters for the classical loss function is studied. In addition, some values of the parameters are considered to be "true" ones and are used to generate synthetic data using the physical model and Gaussian noise, which is then used to study the convergence of the parameters minimizing the presented and the classical loss function, respectively, as the function of the noise parameters.
[+] References (16)
- Gladun, A. V. 2004. Laboratornyy praktikum po obshchey fizike [Laboratory classes on general physics]. Moscow: MFTI. 316 p.
- Rudoy, G. I. 2015. Applying Monte Carlo methods to analysis of nonlinear regression models. Numer. Anal. Appl. 4:344-350.
- Gillard, J. W 2006. An historical overview of linear regression with errors in both variables. Cardiff University School of Mathematics. Technical Report.
- Deming, W. E. 1943. Statistical adjustment of data. New York, NY: Wiley. 216 p.
- Bowden, R. J., D. A. Turkington. 1990. Instrumental vari-ables. Cambridge: Cambridge University Press. 236 p.
- Bekker, P. A. 1986. Comment on identification in the linear errors in variables model. Econometrica 54(1):215- 217.
- Carroll, R. J., D. Ruppert, L.A. Stefanski, and C. M. Crainiceanu. 2006. Measurement error in nonlinear models: A modern perspective. London: Chapman and Hall/CRC. 484 p.
- Jukic, D. 2013. On nonlinear weighted least squares estimation of bass diffusion model. Appl. Math. Comput. 219(14):7891-7900.
- Jukic, D., and D. Markovic. 2010. On nonlinear weight-ed errors-in-variables parameter estimation problem in the three-parameter Weibull model. Appl. Math. Comput. 215(10):3599-3609.
- Kiryati, N., andA.M. Bruckstein. 2000. Heteroscedastic hough transform (HtHT): An efficient method for robust line fitting in the errors in the variables' problem. Comput. Vis. Image Und. 78(1):69-83.
- Boggs, P. T, R. H. Byrd, andR. B. Schnabel. 1987. Astable and efficient algorithm for nonlinear orthogonal distance regression. SIAMJ. Sci. Stat. Comp. 8(6):1052-1078.
- Marquardt, D.W 1963. An algorithm for least-squares estimation of non-linear parameters. J. Soc. Ind. Appl. Math. 11(2):431-441.
- King, D. E. 2009. Dlib-ml: A machine learning toolkit. J. Mach. Learn. Res. 10:1755-1758.
- Aleksandrov, A. Yu., V. A. Dolgikh, I. G. Rudoy, and A. M. Soroka. 1991. Kinetics of a high-pressure electron- beam-excited laser emitting the "yellow" neon line. Sov. J. Quantum Electronics 21(9):933-937.
- Champagne, L. F 1982. Transient optical absorption in the ultraviolet. Applied atomic collision physics. Vol. 3: Gas lasers. Eds. E. W. McDaniel and W. L. Nighan. Amster-dam, Netherlands: Elsevier. 349-386.
- Rudoy, G. I. 2016. Analysis of the stability of nonlin-ear regression models to errors in measured data. Pattern Recognit. Image Anal. 26(3):608-616.
[+] About this article
Title
ON MODIFICATION OF THE MEAN SQUARED ERROR LOSS FUNCTION FOR SOLVING NONLINEAR HETEROSCEDASTIC ERRORS-IN-VARIABLES PROBLEMS
Journal
Informatics and Applications
2017, Volume 11, Issue 2, pp 74-84
Cover Date
2017-06-30
DOI
10.14357/19922264170209
Print ISSN
1992-2264
Publisher
Institute of Informatics Problems, Russian Academy of Sciences
Additional Links
Key words
errors-in-variables models; heteroscedastic errors; symbolic regression; nonlinear regression
Authors
G. I. Rudoy
Author Affiliations
Moscow Institute of Physics and Technology, 9 Institutskiy Per., Dolgoprudny, Moscow Region 141700, Russian Federation
|