A Discrete Data Fitting Models Fusing Genetic Algorithm

Abstract:

To address problems of Least squares method (LSM) fitting curves in application domains, the essay attempts to build a new model by using LMS (Least Median Squares) to analyze the error points, and pretreating the dynamic measuring errors and then getting the fitting curves of testing data. This model is used for electromotor parameters testing which includes load testing and unload testing. Experiments show that the model can erase the influence of outline points, while improving the effects of data curve fitting and reflecting the characteristic of the motor, provide more accurate data fitting curve with small sample data that is in discrete distribution compared with Least squares method.

• Periodical: Advanced Materials Research (Volume 267)
• Pages: 427-432
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.267.427
• Online since: June 2011
• Authors: Tong Rang Fan, Yong Bin Zhao, Lan Wang
• Keywords: Data Fitting, Discrete Data, Genetic Algorithm (GA), Least Median of Squares Regression

References

[1] P. J. Rousseeuw and A. M. Leroy: Robust Regression and Outlier Detection, John Wiley & Sons, New York, (1987).

DOI: 10.1002/0471725382

[2] A. Leroy and P. J. Rousseeuw: Technical Report No. 201, Center for Statistics and O.R., University of Brussels, Belgium, (1984).

[3] P. Huang, W. Wang and J.B. Li: Computer Applications and Software, 2001, 18(4), pp.22-27.

[4] L.M. Desire and L. Kaufman: Analytica Chimica Acta, 1986, 187, pp.171-179.

[5] H.G. Liu, Y. Luo and Y.C. Liu: Journal of Biomedical Engineering, 2006, 23(4) pp.873-877.

[6] C.L. Karr, B. Weck, D.L. Massart and P. Vankeerberghen: Engng. Applic. Artif. Intell. , 1995, 8 (2), pp.177-189.

[7] GB/T 1032-(2005).

[8] IEEE Std 112-(1991).