A simple piecewise cubic spline method for approximation of highly nonlinear data英文文献资料.docVIP

Vol.4, No.1, 79-83 (2012) Natural Science /10.4236/ns.2012.41012 A simple piecewise cubic spline method for approximation of highly nonlinear data Mehdi Zamani Civil Engineering Department, Faculty of Technical and Engineering, Yasouj University, Yasouj, Iran; mahdi@mail.yu.ac.ir Received 14 September 2011; revised 20 October 2011; accepted 30 October 2011 ABSTRACT such situations, if the curve of f(x) passes between and near to the data points, it is more accurate and smoother than when it passes through all the points exactly. Thus, in those cases an approximate analysis is more suitable for predicting and assigning the y values. If the curve governing f(x) were instead to interpolate such data, it would have a sinusoidal or unsmoothed zigzag form that would not be acceptable and would not be usable for data analysis. Most of the publications about this subject are related mainly to linear or simple nonlinear models [1-4]. Some more advanced methods in this field are based on B-spline and Bézier curves [5-7]. These methods are used most for computer graphic. Because of the waviness and the sinusoidal forms of their curves, they are not applicable to engineering problems. Approximation methods are used in the analysis and prediction of data, especially laboratory data, in engineering projects. These methods are usu- ally linear and are obtained by least-square-error approaches. There are many problems in which linear models cannot be applied. Because of that there are logarithmic, exponential and polynomial curve-fitting models. These nonlinear models have a limited application in engineering prob- lems. The variation of most data is such that the nonlinearity cannot be approximated by the ab- ove approaches. These methods are also not ap- plicable when there is a large amount of data. For these reasons, a method of piecewise cubic sp- line approximation has been developed. The mo- del p