# [R] Sum up the values of a function

Petr Pikal petr.pikal at precheza.cz
Fri Jun 10 13:36:34 CEST 2005

```Hi

Your example is not reproducible as we do not know x1.
Do you by chance seek for something like aggregate? If yes see

?aggregate
or
?by

BTW, do you have some trouble with your space bar?

HTH

Petr

On 10 Jun 2005 at 13:08, Stefan Wagner wrote:

> Dear R-Users,
>
> I have to do a maximum-likelihood estimation and have now a problem
> concerning how to sum up my function values in a smart way. I don't
> know how to explain it easyly, so I give you the code and show you
> where my problem is. I have shorten the code a little bit, so that you
> only get the necessary facts:
>
> ws12 <- function (z, i) (1/(1+exp(z[1]*(z[3]-x1[i]-
> z[4]*(m1[i]/n1[i]-0.5))))) ws37 <- function (z, i)
> (1/(1+exp(z[2]*(z[3]-x2[i]- z[5]*(m2[i]/n2[i]-0.5))))) wsAttack12 <-
> function (z,i) (ws12(z,i)*dec1[i]+(1-ws12(z,i))*(1-dec1[i]))
> wsAttack37 <- function (z,i)
> (ws37(z,i)*dec2[i]+(1-ws37(z,i))*(1-dec2[i])) logwsAttack12 <-
> function (z,i) (log(wsAttack12(z,i))) logwsAttack37 <- function (z,i)
> (log(wsAttack37(z,i))) ws12sum <- function (z)
> (logwsAttack12(z,i=1)+logwsAttack12(z,i=2)+logwsAttack12(z,i=3)+logwsA
> ttack12(z,i=4)+logwsAttack12(z,i=5)+logwsAttack12(z,i=6)) ws37sum <-
> function (z)
> (logwsAttack37(z,i=1)+logwsAttack37(z,i=2)+logwsAttack37(z,i=3)+logwsA
> ttack37(z,i=4)+logwsAttack37(z,i=5)+logwsAttack37(z,i=6)+logwsAttack37
> (z,i=7)+logwsAttack37(z,i=8)) wsLOG <- function (z) (ws12sum(z) +
> ws37sum(z)) LogSum <- function (z) (-sum(wsLOG(z))) SP <- c(0.16,
> 0.10, 44, 0.80, 46) out <- nlm (LogSum, p=SP) out
>
> For explanation: x1[i], x2[i], m1[i], m2[i], n1[i], n2[i] are given
> data and z[1:5] are my estimates. My problem is that I have more than
> one session with diffent number of datas so that I am searching for a
> general way of summing up my logwsAttack12 and logwsAttack37. The
> program should recognize how many data are in my table concerning ws12
> and how many concerning ws37 and should, in dependency of z, sum them
> up. I hope you understand my problem and hopefully someone is able to
> solve it. Many thanks for the moment.
>
> Best regards,
>
> Stefan Wagner
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help