a****g
发帖数: 8131 | 1 应该用什么方法? thanks a lot
x- is time
y is response
each line represents a unique subject |
s*******q
发帖数: 310 | |
o****o
发帖数: 8077 | 3 build a quadratic random growth curve, subject=Sub
【在 a****g 的大作中提到】
: 应该用什么方法? thanks a lot
: x- is time
: y is response
: each line represents a unique subject
|
a****g
发帖数: 8131 | 4 thank, will try
【在 s*******q 的大作中提到】
: 用Loess试 下 看看
|
a****g
发帖数: 8131 | 5 thank, will try
【在 o****o 的大作中提到】
: build a quadratic random growth curve, subject=Sub
|
P****D
发帖数: 11146 | 6 Long time ago I used a SAS macro found on the Internet to draw smoothed
curves. I remember it employes LOESS and allows the user to define "smooth
parameter". Just Google loess smooth curve SAS macro... |
a****g
发帖数: 8131 | 7 谢谢
【在 P****D 的大作中提到】
: Long time ago I used a SAS macro found on the Internet to draw smoothed
: curves. I remember it employes LOESS and allows the user to define "smooth
: parameter". Just Google loess smooth curve SAS macro...
|
w*******9
发帖数: 1433 | 8 Don't know how loess could account for the correlation. I think the random
coefficient regression proposed by 2L is more efficient.
【在 P****D 的大作中提到】
: Long time ago I used a SAS macro found on the Internet to draw smoothed
: curves. I remember it employes LOESS and allows the user to define "smooth
: parameter". Just Google loess smooth curve SAS macro...
|
