Displaying 1 result from an estimated 1 matches for "sub_group".
2013 Jan 12
1
Question on broken-line regression: 'segmented' or alternative
...1 <- lm(data[which(data$group==3),3]~data[which(data$group==3),2])
abline(reg_3.1, col="green", lty=1)
#Segmented package to estimate breakpoints and determine slopes for
group2 and 3:
subset_data <- subset(data, subset=data$group!=3)
names(subset_data) <- c("sub_group", "sub_x", "sub_y")
attach(subset_data)
X <- model.matrix(~0+sub_group)*sub_x
time.1 <- X[,1]
time.2 <- X[,2]
time.3 <- X[,3]
olm <- lm(sub_y~0 + sub_group + time.1 + time.2)
os <- segmented(olm, seg.Z= ~ time.1 + time.2,
psi=l...