Statistical data analysis with segmented parabolic regression usage
DOI:
https://doi.org/10.18372/2310-5461.38.12834Keywords:
approximation, segmented (parabolic) regression, optimization of switching point’s abscissa, the best mathematical modelAbstract
The article deals with the problem of the approximation of empirical data using two-segmented parabolic regression. A comparative analysis of this type of approximation with other types of approximating functions (one-segmented polynomials of the second and fourth degrees) was carried out; this analysis allows substantiating the choice of the best mathematical model. The proposed approximation method is based on the introduction of two additional aspects: the use of the Heaviside function for obtaining a general mathematical equation and determining the optimal abscissa for switching point. In order to find the optimum switching point, the criterion for minimizing standard deviation was used. The determination of unknown coefficients for approximating functions was carried out according to the ordinary least squares method. In general, the proposed method of approximation using segmented parabolic regression can be used for the yearly adjustments of the tropical equator location and the corresponding average temperature.
References
Химмельблау Д. Анализ процессов статистическими методами. Перевод с англ. / Д. Химмельблау. – М.: Мир, 1973. – 960 с.
Douglas C. Montgomery, George C. Runger. Applied Statistics and Probability for Engineers, Fours Edition, NJ: John Wiley & Sons, 2007, 768 p.
Mills C. Frederick. Statistical Methods. New York: Pitman Publishing, 1965, 860 p.
Зайцев Г.Н. Математическая статистика в экспериментальной ботанике / Г.Н. Зайцев. – М.: Наука, 1984. – 424 с.
Шараф М.А. Хемометрика. Пер. с англ. / М.А. Шараф, М.А. Иллмэн, Б.Р. Ковальски. – Л.: Химия, 1989. – 272 с.
Reklaitis G.V., Ravindran A., Ragsdell K.M. Engineering optimization. Methods and applications. New York: John Wiley and sons, 1983, 688 p.
