How to run S-Plus on the Mathematics Network 

Leemis
Fall, 2008

S-Plus is an interpreted, object-oriented statistical software package capable
of doing classical and modern statistical computations, data manipulation, and
custom graphics.  The package was originally written at AT&T Bell Labs and is
now owned by Insightful Corp. 

  Entering S-Plus 

    Splus          starts S-Plus 
    Splus -e       starts S-Plus with ability to repeat commands with up arrow
    Splus -g       starts S-Plus in a GUI 

  Exiting S-Plus

    q()            exits S-Plus 

  Executing a file of commands

    source("fn")   executes the commands in filename fn

  Sending output to a file

    sink("fn")     sends output to filename fn rather than the screen


  Comments and help

    # comment      comment for documentation
    help(command)  help on command [type help.start() first on the math network]

  Suspending S-Plus

    ^Z                         suspends S-Plus (fg returns)

  Sample session

    > #                                      1.  WORKING WITH SCALARS 
    >
    > 2 + 2
    >
    > 2 + 3 / 5
    >
    > (2 + 3) / 5
    >
    > sqrt(2) 
    >
    > (sin(8) + exp(4)) / 3
    >
    > gamma(3.5)
    >
    > #                                      2.  WORKING WITH VECTORS
    >
    > x <- c("Hi there", 3.6, 4)
    >
    > x 
    >
    > x <- scan()
    >                # Enter ends input
    > x
    >
    > x <- c(5, -6, 8, 0, -2, 3)
    >
    > x 
    > 
    > x[3]
    >
    > x[3:6]
    >
    > x[c(3, 6)]
    >
    > x[-4]
    >
    > x[-4:-6]
    >
    > x[-c(2,5)]
    >
    > #                                      2A.  Logical subscripts
    >
    > x[x > 0]
    >
    > x[x > mean(x)]
    >
    > y <- x[x > mean(x)]
    >
    > x[x < 0] <- 0
    >
    > name _ c("Caxton", "Fielding", "Stuart", "Brown", "Larson")
    > age _ c(28, 43, 34, 48, 37)
    > salary _ c(32000, 47000, 36000, 62000, 45000)
    > shoe _ c(10.5, 12.0, 9.5, 10.0, 11.0)
    >
    > shoe[name == "Larson"]
    >
    > name[age < 35 & salary < 35000]
    >
    > #                                      2B.  Functions on one vector
    >
    > # sum(x), prod(x), sort(x), min(x), max(x), range(x), length(x), mean(x),
    > # var(x), sin(x), floor(x), ceiling(x), log(x), log10(x), gamma(x),
    > # cumsum(x), cummax(x), ...
    >
    > x <- c(5, -6, 8, 0, -2, 3)
    >
    > max(x)
    >
    > mean(x)
    >
    > sqrt(var(x))
    >
    > sort(x)
    >
    > gamma(x)
    >
    > mode(x)
    >
    > length(x)
    >
    > #                                      2C.  Functions on two vectors
    >
    > y <- c(rep(1, 3), 8:9, 17) 
    > 
    > pmax(x, y)
    >
    > cor(x, y)
    >
    > #                                      2D.  Probability functions 
    >
    > pnorm(-1.5)
    >
    > qnorm(0.98)
    >
    > dnorm(1)
    >
    > rnorm(40)
    >
    > x <- rpois(40, 2.6)
    >
    > pchisq(20, 16)
    >
    > set.seed(6)
    > 
    > rbinom(100, 1, 1 / 2)
    >
    > #                                      2E.  Writing your own functions
    >
    > hyp <- function(a, b) {
    >   temp <- a ^ 2 + b ^ 2
    >   sqrt(temp)
    > }
    >
    > hyp(3, 4)
    >
    > #   QUIZ.  For a vector x with numerical entries.
    > #
    > #   1.  Find the mean of the positive elements of x
    > #   2.  Find the variance of the negative elements of x
    > #   3.  Find the sum of the elements of x, excluding the largest and
    > #       smallest
    > #   4.  Find the subscripts of the positive elements of x
    >
    >
    > #                                      3.  MATRICES 
    >
    > xmat <- matrix(1:20, 4, 5)
    > 
    > length(xmat)
    >
    > dim(xmat)
    > 
    > mode(xmat)
    >
    > ymat <- matrix(runif(20), 5, 4, byrow = T)
    >
    > xmat + matrix(20:1, 4, 5)
    >
    > t(xmat)
    >
    > rbind(xmat, t(ymat))
    >
    > zmat <- xmat %*% ymat
    >
    > sum(diag(zmat))
    >
    > eigen(zmat)
    >
    > xmat[2, 3]
    >
    > xmat[2, ]
    >
    > xmat[, 3]
    >
    > xmat[2:4, ]
    >
    > mean(xmat[1:3, ] > 15)
    >
    > #                                      4.  ARRAYS 
    >
    > a <- array(rpois(60, 3.3), dim = c(3, 4, 5))
    >
    > a[1, 2, 3]
    >
    > a[ , , 5]
    >
    > a[1, ,4]
    >
    > #                                      4A.  The apply function
    >
    > b <- array(1:6, dim = c(2, 3))
    >
    > apply(b, 1, sum)
    >
    > apply(b, 2, sum)
    >
    > apply(b, c(1, 2), sqrt)
    >
    > #                                      5.  Data frames 
    >
    > #                                      6.  Programming 
    >
    > for (i in 1:17) print(c(i, i ^ 2, i ^ 3))
    > 
    > x = 3
    > if (x >= 0) print("x is nonnegative") else print("x is negative")
    >
    >
    > #                                      7.  Built-in data sets 
    >
    > state.x77
    >
    > state.abb
    >
    > state.x77[,8]
    > 
    > sort(state.x77[,8])
    >
    > state.x77[8,]
    >
    > plot(state.x77[,6], state.x77[,7],
    >      xlab = "Percent HS Grads", ylab = "Number days of frost")
    >
    > cu.dimensions
    > bladder
    > cereal.attitude
    >
    > #                                      8.  Built-in statistical procedures
    >
    > x _ c(1986, 1987, 1988, 1989, 1990)
    > y _ c(2.13, 2.35, 2.60, 2.86, 3.23)
    > reg _ lsfit(x, y) 
    > reg
    >
    > fit _ lm(y ~ x)
    > fit
    > summary(fit) 
    >
    > #                                      9.  Graphics 
    >
    > plot(x, y)
    >
    > temp <- rnorm(300)
    >
    > hist(temp)
    >
    > usa()
    >
    > tsplot(rain.nyc1, main = "NYC Rain Data", ylab = "Rain in inches")
    >
    > qqnorm(auto.stats[, "Weight"], ylab = "Weight")
    >
    > contour(switzerland)
    >
    > persp(switzerland)
    >
    > image(voice.five)
    >
    > brush(state.x77)
    >
    > q()
    >
    > # QUIZ SOLUTIONS
    >
    > mean(x[x > 0])
    > var(x[x < 0])
    > sum(sort(x)[-c(1, length(x))])
    > y <- c(1:length(x))
    > y[x > 0]