How to identify sequences within each cluster?

问题:

Using the biofam dataset that comes as part of TraMineR:

library(TraMineR)
data(biofam)
lab <- c("P","L","M","LM","C","LC","LMC","D")
biofam.seq <- seqdef(biofam[,10:25], states=lab)
head(biofam.seq)
     Sequence                                    
1167 P-P-P-P-P-P-P-P-P-LM-LMC-LMC-LMC-LMC-LMC-LMC
514  P-L-L-L-L-L-L-L-L-L-L-LM-LMC-LMC-LMC-LMC    
1013 P-P-P-P-P-P-P-L-L-L-L-L-LM-LMC-LMC-LMC      
275  P-P-P-P-P-L-L-L-L-L-L-L-L-L-L-L             
2580 P-P-P-P-P-L-L-L-L-L-L-L-L-LMC-LMC-LMC       
773  P-P-P-P-P-P-P-P-P-P-P-P-P-P-P-P 

I can perform a cluster analysis:

library(cluster)
couts <- seqsubm(biofam.seq, method = "TRATE")
biofam.om <- seqdist(biofam.seq, method = "OM", indel = 3, sm = couts)
clusterward <- agnes(biofam.om, diss = TRUE, method = "ward")
cluster3 <- cutree(clusterward, k = 3)
cluster3 <- factor(cluster3, labels = c("Type 1", "Type 2", "Type 3"))

However, in this process, the unique id's from biofam.seq have been replaced by a list of numbers 1 through N:

head(cluster3, 10)
[1] Type 1 Type 2 Type 2 Type 2 Type 2 Type 3 Type 3 Type 2 Type 1
[10] Type 2
Levels: Type 1 Type 2 Type 3

Now, I want to know which sequences are within each cluster, so that I can apply other functions to get the mean length, entropy, subsequence, dissimilarity, etc. within each cluster. What I need to do is:

1. Map the old ids to the new ids
2. Insert the sequences in each cluster into separate sequence objects
3. Run the statistics I want on each of the new sequence objects

How can I accomplish 2 and 3 in the list above?

回答1:

The state sequence object for the first cluster, for example, can simply be obtained with

bio1.seq <- biofam.seq[cluster3=="Type 1",]
summary(bio1.seq)

回答2:

I think this will answer your questions. I used the code I found here http://www.bristol.ac.uk/cmm/software/support/workshops/materials/solutions-to-r.pdf to create biofam.seq, since none of what you suggested was working for me.

# create data
library(TraMineR)
data(biofam)
bf.states  <- c("Parent", "Left", "Married", "Left/Married", "Child",
                "Left/Child", "Left/Married/Child", "Divorced")
bf.shortlab <- c("P","L","M","LM","C","LC", "LMC", "D")
biofam.seq  <- seqdef(biofam[, 10:25], states = bf.shortlab,
                                       labels = bf.states)

# cluster
library(cluster)
couts <- seqsubm(biofam.seq, method = "TRATE")
biofam.om <- seqdist(biofam.seq, method = "OM", indel = 3, sm = couts)
clusterward <- agnes(biofam.om, diss = TRUE, method = "ward")
cluster3 <- cutree(clusterward, k = 3)
cluster3 <- factor(cluster3, labels = c("Type 1", "Type 2", "Type 3"))

First, I use split to create a list of indices for each cluster, which I then use in a lapply loop to create a list of sub-sequences from biofam.seq:

# create a list of sequences
idx.list <- split(seq_len(nrow(biofam)), cluster3)
seq.list <- lapply(idx.list, function(idx)biofam.seq[idx, ])

Finally, you can run analytics on each sub-sequence by using lapply or sapply

# compute statistics on each sub-sequence (just an example)
cluster.sizes <- sapply(seq.list, FUN = nrow)

where FUN can be any function you would normally run on a single sequence.