Optim() function error

Question

This is my code.

beta1 = function(a,b,t) { beta(a+(1/t),b) }
beta2 = function(a,b,t) { beta(a+(2/t),b) }

eb11 = function(a,b,t) { beta2(a,b,t)/beta(a,b) }
eb12 = function(a,b,t) { (beta1(a,b,t)-beta2(a,b,t))/beta(a,b) }
eb22 = function(a,b,t) { 1 + (beta2(a,b,t)-2*beta1(a,b,t))/beta(a,b) }

eb11r11 = function(a,b,t) { beta2(a,b,t)*beta(a,b)/beta1(a,b,t)^2 }
eb12r12 = function(a,b,t) { (beta1(a,b,t)-beta2(a,b,t))*beta(a,b)/beta1(a,b,t)/(beta(a,b)-beta1(a,b,t)) }
eb22r22 = function(a,b,t) { (beta(a,b)^2 + (beta2(a,b,t)-2*beta1(a,b,t))*beta(a,b))/(beta(a,b)-beta1(a,b,t))^2 }

gbetloglik = function(a,b,t) {
  loglik = n1*log(eb11r11(a,b,t)) + n2*log(eb12r12(a,b,t)) + n3*log(eb22r22(a,b,t))
  return(-loglik)
}

abt = optim(c(0.5,0.5,1),gbetloglik,lower=c(0.001,0.001,0.001),method="L-BFGS-B")$par

What I'd like to do is to find a,b, and t that maximize 'gbetloglik' function. But I got this error.

Error in 2/t : 't' is missing

It seems that the third argument of function 'beta2' is missing. When I enter three numbers directly in gbetloglik function, it works well. The problem occurs only in optim() function. Does anyone have any idea?

crogg01 · Accepted Answer · 2014-02-14 22:18:14Z

2

It looks like you are misinterpreting the first argument of the optim function. The first argument simply supplies initial values for the 1 arguments being optimized. In your case this is supplying 3 initial guesses for one of the arguments to gbetloglik. This call will work:

abt = optim(0.5,gbetloglik,lower=c(0.001,0.001,0.001),method="L-BFGS-B", b=0.5, t= 0.5)$par

but won't optimize across all three arguments, it will simply optimize a given b and t. To optimize across all arguments you will need to install an external package from here. Here is an example from nlmrt:

ydat  =  c(6.308, 6.94, 9.638, 12.866, 17.069, 23.192, 31.443, 37.558, 51.156, 64.948, 77.995, 91.972)
tdat  =  seq_along(ydat)
start1  =  c(b1=1, b2=1, b3=1)
eunsc  =   y ~ b1/(1+b2*exp(-b3*tt))
anlxb1g =try(nlxb(eunsc, start=start1, trace=FALSE, data=data.frame(y=ydat, tt=tdat)))
print(anlxb1g)
anlxb1g$coefficients

edited Feb 14, 2014 at 22:18

answered Feb 14, 2014 at 21:12

crogg01

2,52618 silver badges35 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Optim() function error

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related