Newton's Method in R

Question

I have an issue when trying to implement the code for Newton's Method for finding the value of the square root (using iterations). I'm trying to get the function to stop printing the values once a certain accuracy is reached, but I can't seem to get this working. Below is my code.

MySqrt <- function (x, eps = 1e-6, itmax = 100, verbose = TRUE){
  i <- 1
  myvector <- integer(0)
  GUESS <- readline(prompt="Enter your guess: ")
  GUESS <- as.integer(GUESS)
  while(i <= itmax){
      GUESS <- (GUESS + (x/GUESS)) * 0.5
      myvector <- c(myvector, GUESS)
      if (abs(GUESS-x) < eps) break
      i <- i + 1
  }

  myvector

Why won't the if-statement work?

Answer 1

This should work:

MySqrt <- function (x, eps = 1e-6, itmax = 100, verbose = TRUE){
  i <- 1
  myvector <- vector(mode='numeric',itmax)  ## better to allocate memory
  GUESS <- readline(prompt="Enter your guess: ")
  GUESS <- as.numeric(GUESS)
  myvector[i] <- GUESS
  while(i <= itmax){
    GUESS <- (GUESS + (x/GUESS)) * 0.5
    if (abs(GUESS-myvector[i]) < eps) break
    i <- i + 1
    myvector[i] <-  GUESS
  }
  myvector[seq(i)]
}

MySqrt(2)
Enter your guess: 1.4
[1] 1.400000 1.414286 1.414214

Answer 2

UPDATE:

Please see @RichieCotton's comment to @agstudy's answer. I agree with Richie, and in fact it makes more sense to use @agstudy's approach.

---

Original answer:

Your function is fine, your math is off.
GUESS and x should not (necessarilly) be close, but GUESS * GUESS and x should be.

MySqrt <- function (x, eps = 1e-6, itmax = 100, verbose = TRUE){
  i <- 1
  myvector <- integer(0)
  GUESS <- readline(prompt="Enter your guess: ")
  GUESS <- as.integer(GUESS)
  while(i <= itmax){
      GUESS <- (GUESS + (x/GUESS)) * 0.5
      myvector <- c(myvector, GUESS)
      browser(expr={i == 10 || abs(GUESS-x) < eps})
      if (abs((GUESS*GUESS)-x) < eps) break    ###  <~~~~  SEE HERE
      i <- i + 1
  }

  myvector
}