0%

Multidimensional Scaling in R

I met this problem when I was analysing teachers’ PCK concept maps. I wanted to divide them into several clusters that shared the similar structures(nodes and links).

One approch is plotting them in a 2d picture to see if there is any patterns. But before that, you have to do the Muultidimensional Scaling to get your (x,y) from your Distance Matrix.

Distance Matrix
Distance Matrix
How I got this matrix. Well, I just count them using my data. Of course, if can deside how you count them, which counts how much score, or you simply give them a number.


# 1) MDS 'cmdscale'
dm <- read.table("distance-matrix.txt")
mds1 = cmdscale(dm, k = 2)

# plot
plot(mds1[,1], mds1[,2], type = "n", xlab = "", ylab = "", axes = FALSE, main = "cmdscale (stats)")
text(mds1[,1], mds1[,2], labels(x)[[1]], cex = 0.9, xpd = TRUE)

This the x y data I got:
Distance Matrix

I got this picture after plotting it out.
Distance Matrix

Reference:
7 Functions to do Metric Multidimensional Scaling in R | Gaston Sanchez

感谢您的鼓励和打赏!