| 
|
Readme file
SERIES
B
 Statistical
Methodology
On distribution-weighted partial least squares with diverging number of highly correlated predictors, by Li-Ping Zhu and Li-Xing Zhu, pages 525–548
Estimate the direction in the model Y = G(\beta^T X, e) by distribution-weighted partial least quares
Usage: DPLS = DWPLS(x, y, K)
Input:
x -- the observation predictors matrix of size [n, p] where n is
the sample size and p is the dimension of predictors
y -- continuous response vector of size [n, 1];
K -- the dimension of Krylov space. In applications K is often very
small. We can usually choose K = 1, 2 or 3. One can also refer to
Section 3.1 (15) for a BIC type criterion to estimate the dimension.
Output:
DPLS -- the estimated direction by DWPLS
Two examples are given in the file Example_Data.txt.
Li-Ping Zhu
School of Finance and Statistics
East China Normal University
Shanghai
People's Republic of China
E-mail: lpzhu@stat.ecnu.edu.cn
|
Journals
SERIES
A
Statistics
in Society
SERIES
B
Statistical
Methodology
SERIES C
Applied Statistics
SERIES D
The
Statistician
|
