Given a scatterplot of \((x_i, y_i)\), \(i = 1,\ldots,n\), where \(\bold{x}\) is the predictor and \(\bold{y}\) is the response which could be Gaussian, Poisson and binomial. The underlying mean curve in this scatterplot is denoted as \(\bold{f_m(x)}\), where \(\bold{f_m(x)}\) is a smooth curve with a change-point \(m\) which falls in one of the three categories:
a mode in a unimodal smooth curve
an inflection point in a convex-concave (concave-convex) smooth curve
a jump point in an otherwise smooth curve.
Given the category of the change-point specified by the user, the main routine "changept" estimates the mean curve and the change-point as well using shape-restricted B-splines.
See changept for more details.
Xiyue Liao and Mary C. Meyer
Maintainer: Xiyue Liao <xiyue@rams.colostate.edu>
| Package: | ShapeChange |
| Type: | Package |
| Version: | 1.3 |
| Date: | 2016-03-19 |
| License: | GPL (>= 2) |