edr.n {censorSIR}R Documentation

get e.d.r. directions

Description

take a csir(result of cen.sir) or ds(result of double.slice) object as input, return the effective dimension reduction directions.

Usage

edr.n(object, n)
edr.percent(object, percent)

Arguments

object a variable of class csir(result of cen.sir) or ds(result of double.slice)
n the number of directions to return
percent the minimum percentage of information that the returned directions should explain, real number between 0 and 1

Details

edr.n simply return the first n eigen-vectors of an csir or ds object. edr.percent simply return the first n eigen-vectors that explain at least percent information. To decide how many directions should be used, users can read the print-out of a csir or a ds object, where result of Chi-Square test and size of eigen values are shown.

Value

a matrix, each column is one eigen-vector to be used.

Note

These two functions: edr.n and edr.percent do not decide how many directions should be counted as e.d.r. directions. There information depend on users input.

Author(s)

Wei Sun sunwei@stat.ucla.edu

References

Ker-chau Li, Jane-ling Wang and Chun-Hou Chen (1999) Dimension reduction for censored regression data. The Annals of Statistics 27, 1-23.

See Also

double.slice cen.sir


[Package censorSIR version 1.0 Index]