Nonparametric Kernel Smoothing Methods. ThesmLibrary in Xlisp-Stat

Luca Scrucca
2001 Journal of Statistical Software  
In this paper we describe the Xlisp-Stat version of the sm library, a software for applying nonparametric kernel smoothing methods. The original version of the sm library was written by Bowman and Azzalini in S-Plus, and it is documented in their book Applied Smoothing Techniques for Data Analysis (1997). This is also the main reference for a complete description of the statistical methods implemented. The sm library provides kernel smoothing methods for obtaining nonparametric estimates of
more » ... ity functions and regression curves for different data structures. Smoothing techniques may be employed as a descriptive graphical tool for exploratory data analysis. Furthermore, they can also serve for inferential purposes as, for instance, when a nonparametric estimate is used for checking a proposed parametric model. The Xlisp-Stat version includes some extensions to the original sm library, mainly in the area of local likelihood estimation for generalized linear models. The Xlisp-Stat version of the sm library has been written following an object-oriented approach. This should allow experienced Xlisp-Stat users to implement easily their own methods and new research ideas into the built-in prototypes.
doi:10.18637/jss.v006.i07 fatcat:fpl6ljr74fadrnwe5t3lxrolia