CODEWITHSUNDEEP
r
omba
omba

Reputation: 1

confidence ellipse came out as y shape

I drew the confidence ellipse for mu1 and mu 2 i used following code based on means and variance inverse I got but the ellipse came out as Y shape:

enter image description here

library(car)
library(SIBER)
ellipse<-function(mu1,mu2)
{
  30*1.155461e-05*(1860.500-mu1)^2+ 30*4.11493e-07*(8354.133-mu2)^2
  - 60*1.98504e-06*(1860.500-mu1)*(8354.133-mu2)
}
mu1<-seq(1650,2100,50)
mu2<-seq(7000,9500,50)
z <- t(sapply(mu1,ellipse,mu2))
contour(mu1,mu2,z, 
        levels=6.91, 
        drawlabels=T, 
        axes=T,
        frame.plot=T, 
        xlab="mu1 (stf)",
        ylab="mu2 (ben strnegth)",
        main = "A 95% confidence ellipse for mu = (mu1,mu2)")
points(1860.5,8354.133)
segments(0,8354.133,1860.500,8354.133)
segments(1860.500,0,1860.500,8354.133)

Upvotes: 0

Views: 51

Answers (1)

Ben Bolker
Ben Bolker

Reputation: 226172

A surprisingly subtle typo. In your function definition,

ellipse<-function(mu1,mu2) {
  30*1.155461e-05*(1860.500-mu1)^2+ 30*4.11493e-07*(8354.133-mu2)^2
  - 60*1.98504e-06*(1860.500-mu1)*(8354.133-mu2)
}

you broke the line before the minus sign; the first line is a complete R expression, so R computes it (and discards the result since you didn't assign it to a value), and then computes the second line as a second expression and returns it. In other words, your function is only returning the cross-term (60*1.98504e-06 ...).

Possible solutions:

  • put parentheses around the whole expression
  • move the minus sign to the end of the first line, so that R recognizes that the expression isn't finished yet

You may find the ellipse package or the ellipse function in the car package useful ...

Upvotes: 2

Related Questions