function call in R what does p in below code mean

Question

I am new to R and trying to understand this function call below:

mle_mos <- function (n, m1, m2, x, x1, x2, initial, iters)  
    mvnewton (function (p) log_likelihood_gradient(p,n,m1,m2,x,x1,x2),   
            function (p) -fisher_information(p,n,m1,m2,x,x1,x2),   
            initial, iters)  


mvnewton <- function (f, g, x, n)  
{  
    print(f)
    if (n < 1)  
    stop("invalid number of iterations")  
  
    for (i in 1:n) {  
        cat("\nAt iteration",i,":\n\n")  
        cat("x =",x,"\n")  
        cat("f(x) =",f(x),"\n")  
        cat("g(x) =\n")  
        print(g(x));  
        x <- x - solve(g(x),f(x))  # computes inverse of g(x) times   
        f(x)  
    }  
    x  
}

mvnnewton takes
(function (p) log_likelihood_gradient(p,n,m1,m2,x,x1,x2)
as the first parameter. Where is p getting it's value from? Where is it defined? Can someone explain where p is getting a value to be passed into log_likelihood_gradient as a parameter? I have been googling for the past 2 days and reading up a lot of stuff but am still not clear that I understand this properly.

Your code is not clear. Please show a small reproducible example — akrun
– akrun, Commented Feb 20, 2022 at 16:31
function (p) log_likelihood_gradient(p,n,m1,m2,x,x1,x2) is a definition of a function. It will get the value from the mvnewton function, for which it is an argument. — user2554330
– user2554330, Commented Feb 20, 2022 at 16:32
updated code, thanks user2554330, but I still don't understand — kam amar
– kam amar, Commented Feb 20, 2022 at 17:15

G. Grothendieck · Accepted Answer · 2022-02-20 18:45:14Z

1

Arguments

If we have a function fun

fun <- function(z) z+1

and then we call it

fun(1)

1 is said to be the actual argument and z is said to be the formal argument. The formal argument takes on the value of the actual argument so z takes on the value 1. (There are some technicalities which we have glossed over but these do not affect the situation in the question.)

mvnewton

When mvnewton is called the caller passes to it four actual arguments:

function (p) log_likelihood_gradient(p,n,m1,m2,x,x1,x2)

function (p) -fisher_information(p,n,m1,m2,x,x1,x2)

initial

iters

Note that the first two arguments are functions. It does not pass the result of calling the functions but passes the functions themselves.

Now within mvnewton these 4 actual arguments correspond to the formal arguments f, g, x and n so when mvnewton refers to f it is really referring to

function (p) log_likelihood_gradient(p,n,m1,m2,x,x1,x2)

Also within mvnewton the function f is called with actual argument x and x corresponds to formal argument p within f (and similarly for g).

Example

For example, suppose we call

f <- function(x) x^2
g <- function(x) 2*x
mvnewton(f, g, 1, 20)

Then x is 1 and the first time f(x) and g(x) are called within mvnewton they are called as f(1) and g(1) and within f and g the formal argument p takes the value 1. mvnewton then updates x to a new value, calls f and g again with the new value of x and so on.

edited Feb 20, 2022 at 18:45

answered Feb 20, 2022 at 17:38

G. Grothendieck

273k18 gold badges221 silver badges365 bronze badges

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

2 Comments

kam amar Over a year ago

What is the difference between the p in function(p) and log_likelihood_gradient(p,n,m1,m2,x,x1,x2)

G. Grothendieck Over a year ago

In function (p) log_likelihood_gradient(p,n,m1,m2,x,x1,x2) the first p is a formal argument and it is passed as an actual argument in the call to log_likelihood_gradient so log_likelihood_gradient winds up using whatever value was passed to function (p) log_likelihood_gradient(p,n,m1,m2,x,x1,x2) when the latter is called.

Collectives™ on Stack Overflow

function call in R what does p in below code mean

1 Answer 1

Arguments

mvnewton

Example

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Arguments

mvnewton

Example

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related