Getting values from indicator variables in R

问题

Firstly I had to model the coordinates of two points in a 20m X 30m rectangle. Obviously these points follow a uniform distribution so this is the first part of my code:

    X1 <- runif(1,0,20)
    X2 <- runif(1,0,30)
    point1 <- c(X1,X2)
    point1

I used the same code for the second point ('point2') but replaced the X1 and X2 with Y1 and Y2 respectively.

I then had to find the distance between the two points:

    distance <- sqrt(((X1-Y1)^2)+((X2-Y2)^2))

Now if I define A as the event where the points are within 5m to 10m of each other, I need to find the indicator variable of this event. This is what I got to, but I'm not sure if it's right:

    x=0
    if (distance < 10 & distance > 5)
    {
     x=1
    }
    Z <- c(distance,x)

If I were to repeat these commands 1000 times, how would I store the values in each simulation and find the min and max separation values in the 1000 reps?

回答1:

At one point you use X1 and X2 and later refer to Y1 which has not been defined. I'm pretty sure you want to use:

 points <- cbind(X1, X2)

Context suggests that you wanted to keep the X1 and X2 values in a "parallel" arrangement and defining a matrix rather than a non-dimensioned object will accomplish that.

Last question answered using R's matrix operations:

points1 <- matrix( runif(2000), ncol=2)
points1 <- matrix( runif(2000), ncol=2)
dists <- rowSums( (points1-points2)^2 )
Z <- dists <10 & dists >5

回答2:

Close. Use ifelse or simply a vectorised form to define your event. For 1000 samples all you need to do is use runif to generate 1000 samples. Also there is no need to c(X1 , X2), you can just call them as vectors in your distance calculation...

#Just make 1000 calls to runif 
X1 <- runif(1000,0,20)
X2 <- runif(1000,0,30)
Y1 <- runif(1000,0,20)
Y2 <- runif(1000,0,30)
distance <- sqrt(((X1-Y1)^2)+((X2-Y2)^2))

head(distance)
#[1]  9.050522 19.512849 10.413407  7.736564  2.742174 13.729397

# gives 1 if points are within 5 to 10m of each other
event <- ifelse ( distance >= 5 & distance <= 10 , 1 , 0 )

#Or even better, from @GavinSimpson's comment just use a vectorised form (we use as.nuemric to turn TRUE / FALSE into 1 and 0, but you could leave that out if you wish)
event <- as.numeric( distance >= 5 & distance <= 10 )

head( event )
#[1] 1 0 0 1 0 0

# And the minimum and maximum sepration distance of those events
min(distance[distance >= 5 & distance <= 10])
#[1] 5.017296
max(distance[distance >= 5 & distance <= 10])
#[1] 9.989868

来源：https://stackoverflow.com/questions/15526782/getting-values-from-indicator-variables-in-r

标签

statistics