Nico Met' ; 'R help'
Cc:
Sent: Monday, July 15, 2013 1:33 PM
Subject: Re: [R] t-test across columns
This may be close to what you want:
> t.val <- by(x, x$cat, function(y) if (min(table(y$name)>1)) {
+ t.test(val~name, y)})
> t.out <- do.call(rbind, sapply(t.val, f
This may be close to what you want:
> t.val <- by(x, x$cat, function(y) if (min(table(y$name)>1)) {
+ t.test(val~name, y)})
> t.out <- do.call(rbind, sapply(t.val, function(y) c(y$statistic,
+ p.value=y$p.value)))
> t.out
t p.value
p178266580 -0.1156475 0.91440
Hi,
Not sure about the format for the 2nd part.
df1<- ##data
library(plyr)
df2<-ddply(df1,.(name,cat),summarize,
cbind(t.test(val,df1$val)$statistic,t.test(val,df1$val)$p.value))
df3<-cbind(df2[,1:2],data.frame(df2[,3]))
colnames(df3)[3:4]<- c("t-val","p.val")
library(reshape2)
df3m<- melt(df3
3 matches
Mail list logo
