Hello everybody, I would like to perform some treatments on a group. Each element of my group is a 2D array. They all have the same dimensions. I would like to know if I can find in R some functions like "dist" (base package), "hclust" and "kmeans" (mva package) but in order to use 2D arrays more than vectors. Maybe my question doesn't have sense for the R context but it would be really great for my job :-) Many thank in advance for any help Li?m TRAN -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
2D fellows
3 messages · TRAN Liem FTM/Men, Winfried Theis
Dear Liem!
On 15-May-00 TRAN Liem FTM/Men wrote:
Hello everybody, I would like to perform some treatments on a group. Each element of my group is a 2D array. They all have the same dimensions.
I don't know what you mean exactly by a "2D-array", I would think first of something you could easily convert to a matrix, or a data frame! (I just had a look on "Notes on R" and the description of arrays there confirms my belief about the convertion.)
I would like to know if I can find in R some functions like "dist" (base package), "hclust" and "kmeans" (mva package) but in order to use 2D arrays more than vectors.
Well, if you could convert your arrays into vectors (via matrix or data frame if you want to) and collect them in a data frame you could use these functions. If you need a special distance you could program a function in which you compute the pairwise distances and write them to a dissimilarity matrix as in the package cluster. For a description see the documentation there (or just type daisy after loading the package cluster). With this you could use the functions diana for divisive hierarchical clustering, agnes for agglomerative nesting or for k-means pam. I hope this helps, yours Winfried Theis ---------------------------------- E-Mail: Winfried Theis <theis at statistik.uni-dortmund.de> Date: 16-May-00 Tel.: +49-231-755-5903 Room: Mathematikgebäude, M832 Dipl.-Math. Winfried Theis, Graduiertenkolleg "Angewandte Statistik", Universität Dortmund, 44221 Dortmund ---------------------------------- -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Hello Winfried
I don't know what you mean exactly by a "2D-array", I would think first of something you could easily convert to a matrix, or a data frame! (I just had a look on "Notes on R" and the description of arrays there confirms my belief about the convertion.)
Indeed, my objects can be considered as matrix or data frame
Well, if you could convert your arrays into vectors (via matrix or data frame if you want to) and collect them in a data frame you could use these functions.
It would give me (for example with "dist") the (relative) distances between my collection of rows. My problem (I realize I wasn't precise enough) is to calulate, for example, the distance between two matrices and not between two rows. If I consider 2 matrices A and B with N rows each (and M columns), I could use dist on each pair of rows (one for A, one for B), and compute N time "dist" on 2 x M matrices. And I could after calculate a distance with the N distances but I find it a bit heavy, and I don't know how to perform an object which would be usable as input for, let's say, "hclust".
If you need a special distance you could program a function in which you compute the pairwise distances and write them to a dissimilarity matrix as in the package cluster.
That is typically my problem with R: How can I write the right object like a dissimilarity matrix if I need one? I'm not used enough with the "statistical" objects and not used enough with R to succeed writing directly the objects in the right "R-shape". That's why I would like to find already "R-shaped" objects.
For a description see the documentation there (or just type daisy after loading the package cluster). With this you could use the functions diana for divisive hierarchical clustering, agnes for agglomerative nesting or for k-means pam.
I just installed the cluster package among others. I hope to find what I'm looking for. Otherwise, I'll be back ;-)
I hope this helps,
It does. Many thanks Li?m -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._