hello,
i'm new in writting functions in R. I read the R-help manuals but still
couldn't get it right.
The problem is:
I have data on 16 categorical variables for 335 individuals. the data are
lines for individuals and columns for variables.
I want to calculate the correlation between two variables using a formula
given by Agresti for ordinal categorical variables.
I have to calcule the value for each pair of variables and save the result in
a vector...
I wrote the following code
X<-as.matrix(X)
n=335
for (i in 1:15){
for (j in i+1:16){
Tab<-table(X[,i],X[,j])
u<-1:nrow(Tab)
v<-1:ncol(Tab)
ml<-margin.table(Tab,1)
mc<-margin.table(Tab,2)
r1<-u*Tab
r2<-sum(v*t(r1))
r3<-sum(u*ml)*(sum(v*mc)/n
r5<-sum((v^2)*mc)-((sum(v*mc))^2)/n
r4<-sum((u^2)*ml)-((sum(u*ml))^2)/n
r<-(r2-r3)/sqrt(r4*r5)
r
m=(n-1)*r^2
}
}
The correlation is r.I have a result, but there is a warning regarding r5
about sintax.
Plus, it doesn??t perform the loops and I don't know how to save the results
in
a vector.
thanks for any help
Assun????o
__________________________________________________________
>From: assuncao.senra at portugalmail.com >To: R help <r-help at stat.math.ethz.ch> >Subject: [R] help in building a function >Date: Sat, 11 Sep 2004 14:22:13 +0100 > >hello, >i'm new in writting functions in R. I read the R-help manuals but still >couldn't get it right. >The problem is: >I have data on 16 categorical variables for 335 individuals. the data are >lines for individuals and columns for variables. >I want to calculate the correlation between two variables using a formula >given by Agresti for ordinal categorical variables. >I have to calcule the value for each pair of variables and save the result >in >a vector... >I wrote the following code > >X<-as.matrix(X) > >n=335 >for (i in 1:15){ >for (j in i+1:16){note i+1:16 stands for i+(1:16)>Tab<-table(X[,i],X[,j]) >u<-1:nrow(Tab) >v<-1:ncol(Tab) > >ml<-margin.table(Tab,1) >mc<-margin.table(Tab,2) > >r1<-u*Tab >r2<-sum(v*t(r1))t is useless here, also you did not use t at other place> >r3<-sum(u*ml)*(sum(v*mc)/n > >r5<-sum((v^2)*mc)-((sum(v*mc))^2)/n > >r4<-sum((u^2)*ml)-((sum(u*ml))^2)/n >r<-(r2-r3)/sqrt(r4*r5) > >r >m=(n-1)*r^2 >} >} >The correlation is r.I have a result, but there is a warning regarding r5 >about sintax. >Plus, it doesn´t perform the loops and I don't know how to save the results >in >a vector. >thanks for any help > >Assunção > > > >__________________________________________________________ > >______________________________________________ >R-help at stat.math.ethz.ch mailing list >stat.ethz.ch/mailman/listinfo/r-help >PLEASE do read the posting guide! >R-project.org/posting-guide.html