7 Jan
2005
7 Jan
'05
6:28 a.m.
Hi!
I'm still a bit (quite) confused about the use of zelig, method "mlogit"
trying to work 'through the given example (but with my own data) I would like to use
the ternaryplot function from vcd, but plot also the intermediary categories
I have got the object z.out a
x.weak<-setx(z.out, v1=1)
x.int1<-setx(z.out, v1=2)
x.int2<-setx(z.out, v1=3)
x.strong<-setx(z.out, v1=4)
From the example:
s.out <- sim(z.out, x =
x.strong, x1 = x.weak)
ev.weak <- s.out$qi$ev + s.out$qi$fd
ternaryplot(s.out$qi$ev, pch = 20, col = "gray50", main = "Y and
v1");
ternarypoints(ev.weak, pch = 20, col = "black")
- works fine!
how do I get the intermediary points on the plot? does teh following make sense?
ternaryplot(s.out$qi$ev, pch = ".", col = "gray50", main =
"dependances OPARS et DMST manger");ternarypoints(ev.weak, pch = ".",
col = "black")
ev.weak1 <- s.out1$qi$ev + s.out1$qi$fd
ev.weak21 <- s.out2$qi$ev + s.out2$qi$fd
ternarypoints(ev.weak2, pch = ".", col = "red")
ternarypoints(ev.weak1, pch = ".", col = "blue")
Now, moreover if my response has more than 3 levels have you a suggestion of how to plot
the regression results?
Anne
PS I'm sorry but I'm new to zelig AND polytomous logistic regression!
----------------------------------------------------
Anne Piotet
Tel: +41 79 359 83 32 (mobile)
Email: anne.piotet(a)m-td.com
---------------------------------------------------
M-TD Modelling and Technology Development
PSE-C
CH-1015 Lausanne
Switzerland
Tel: +41 21 693 83 98
Fax: +41 21 646 41 33
--------------------------------------------------
7 Jan
7 Jan
8:37 a.m.
ternary diagrams only work for 3 outcomes at a time. more than that and
you add dimensions. if you want to do graphics still, you can try adding
color. otherwise, i'd do it with tables of predicted values given
specified values of the explanatory variables.
Best of luck with your research,
Gary King
: Gary King, King(a)Harvard.Edu http://GKing.Harvard.Edu :
: Center for Basic Research Direct (617) 495-2027 :
: in the Social Sciences Assistant (617) 495-9271 :
: 34 Kirkland Street, Rm. 2 HU-MIT DC (617) 495-4734 :
: Harvard U, Cambridge, MA 02138 eFax (617) 812-8581 :
On Fri, 7 Jan 2005, Anne wrote:
Hi!
I'm still a bit (quite) confused about the use of zelig, method "mlogit"
trying to work 'through the given example (but with my own data) I would like to use
the ternaryplot function from vcd, but plot also the intermediary categories
I have got the object z.out a
x.weak<-setx(z.out, v1=1)
x.int1<-setx(z.out, v1=2)
x.int2<-setx(z.out, v1=3)
x.strong<-setx(z.out, v1=4)
From the example:
s.out <- sim(z.out, x = x.strong, x1 = x.weak)
ev.weak <- s.out$qi$ev + s.out$qi$fd
ternaryplot(s.out$qi$ev, pch = 20, col = "gray50", main = "Y and
v1");
ternarypoints(ev.weak, pch = 20, col = "black")
- works fine!
how do I get the intermediary points on the plot? does teh following make sense?
ternaryplot(s.out$qi$ev, pch = ".", col = "gray50", main =
"dependances OPARS et DMST manger");ternarypoints(ev.weak, pch = ".",
col = "black")
ev.weak1 <- s.out1$qi$ev + s.out1$qi$fd
ev.weak21 <- s.out2$qi$ev + s.out2$qi$fd
ternarypoints(ev.weak2, pch = ".", col = "red")
ternarypoints(ev.weak1, pch = ".", col = "blue")
Now, moreover if my response has more than 3 levels have you a suggestion of how to plot
the regression results?
Anne
PS I'm sorry but I'm new to zelig AND polytomous logistic regression!
----------------------------------------------------
Anne Piotet
Tel: +41 79 359 83 32 (mobile)
Email: anne.piotet(a)m-td.com
---------------------------------------------------
M-TD Modelling and Technology Development
PSE-C
CH-1015 Lausanne
Switzerland
Tel: +41 21 693 83 98
Fax: +41 21 646 41 33
--------------------------------------------------
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)latte.harvard.edu
[un]subscribe Options: http://lists.hmdc.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
2:13 p.m.
Hi. One thing you can do is reduce the four dimensions to three, then use
ternary plots or some other technique. The cmdscale routine in the stats
library does this nicely. It requires, however, that you create a set of
distances, or dissimilarities, among your observations on the dimensions.
You can do this by using, for example, a three dimensional representation of
four dimensions, where one of the dimensions is the vector length, the other
three are x,y,z. Then calculate some distance between the end points of the
vectors, using the resultant matrix as input into cmdscale. cmdscale will
give you the best ('in a least squares sense') representation of the 4
dimensions in 2 or three.
Working paper 35 at www.csss.washington.edu gives an example (also used in
Political Analysis 12(2) 2004). Let me know if a code clip will help.
Good luck.
Mike Ward
From: "Anne" <anne.piotet(a)urbanet.ch>
To: "zelig" <zelig(a)latte.harvard.edu>
Subject: [zelig] help with poltting polytomous logistic regression
Date: Fri, 7 Jan 2005 12:28:20 +0100
Hi!
I'm still a bit (quite) confused about the use of zelig, method "mlogit"
trying to work 'through the given example (but with my own data) I would
like to use the ternaryplot function from vcd, but plot also the
intermediary categories
I have got the object z.out a
x.weak<-setx(z.out, v1=1)
x.int1<-setx(z.out, v1=2)
x.int2<-setx(z.out, v1=3)
x.strong<-setx(z.out, v1=4)
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)latte.harvard.edu
[un]subscribe Options: http://lists.hmdc.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
From the example:
s.out <- sim(z.out, x =
x.strong, x1 = x.weak)
ev.weak <- s.out$qi$ev + s.out$qi$fd
ternaryplot(s.out$qi$ev, pch = 20, col = "gray50", main = "Y and
v1");
ternarypoints(ev.weak, pch = 20, col = "black")
- works fine!
how do I get the intermediary points on the plot? does teh following make
sense?
ternaryplot(s.out$qi$ev, pch = ".", col = "gray50", main =
"dependances
OPARS et DMST manger");ternarypoints(ev.weak, pch = ".", col =
"black")
ev.weak1 <- s.out1$qi$ev + s.out1$qi$fd
ev.weak21 <- s.out2$qi$ev + s.out2$qi$fd
ternarypoints(ev.weak2, pch = ".", col = "red")
ternarypoints(ev.weak1, pch = ".", col = "blue")
Now, moreover if my response has more than 3 levels have you a suggestion
of how to plot the regression results?
Anne
PS I'm sorry but I'm new to zelig AND polytomous logistic regression!
----------------------------------------------------
Anne Piotet
Tel: +41 79 359 83 32 (mobile)
Email: anne.piotet(a)m-td.com
---------------------------------------------------
M-TD Modelling and Technology Development
PSE-C
CH-1015 Lausanne
Switzerland
Tel: +41 21 693 83 98
Fax: +41 21 646 41 33
--------------------------------------------------
7071
days inactive
7071
days old
2 comments
3 participants
participants (3)
-
Anne -
Gary King -
Michael Ward