Bassingthwaighte J B, Chan I S, Goldstein A A
Center for Bioengineering, University of Washington, Seattle 98195.
Comput Biomed Res. 1988 Apr;21(2):192-202. doi: 10.1016/0010-4809(88)90027-4.
A mathematical function has been developed for approximating unimodal functions, particularly those which are non-Gaussian, skewed, and incomplete. It is useful as an alternative to cubic splines in smoothing noisy experimental data. Particular applications are to indicator-dilution curves and probability density functions of varied form. In its Fortran implementation, SMOEX, it is computationally inexpensive compared to standard cubic spline smoothing routines and requires less storage to preserve the smoothed function for retrieval or interpolation.
已开发出一种数学函数，用于逼近单峰函数，特别是那些非高斯、偏态且不完整的函数。它可作为三次样条插值的替代方法，用于平滑有噪声的实验数据。其具体应用包括示踪剂稀释曲线和各种形式的概率密度函数。在其Fortran实现SMOEX中，与标准三次样条平滑例程相比，它计算成本较低，并且为保存平滑函数以供检索或插值所需的存储空间较少。
