R arulesSequences - which frequent sequences are present in a transaction?

Question

When doing frequent sequence mining, one typically wants to do the following tasks:
1. Find sequential patterns (frequent sequences).
2. Find out which sequential patterns apply to a transaction. I.e.: given a transaction, which of all of the frequent sequences found is present?

I'm having trouble doing the latter.

Using R, I am applying the cspade-algorithm from the arulesSequences package on the following toy dataset:

data <- data.frame(id = 1:10,
                   transaction = c("A B B A",
                                   "A B C B D C B B B F A",
                                   "A A B",
                                   "B A B A",
                                   "A B B B B",
                                   "A A A B",
                                   "A B B A B B",
                                   "E F F A C B D A B C D E",
                                   "A B B A B",
                                   "A B"))

Then I split the data using the str_split function from package stringr:

data_for_fseq_mining <- str_split(string = data$transaction, pattern = " ")

Use identifiers to uniquely name the list elements in 'data_for_fseq_mining'. This is a prerequisite for using the function 'as.transactions' as shown below.

names(data_for_fseq_mining) <- data$id

In order to convert this kind of data to a dataset of class 'transactions' I use the following function as.transactions from https://github.com/cran/clickstream/blob/master/R/Clickstream.r.

data_for_fseq_mining_trans <- as.transactions(clickstreamList = data_for_fseq_mining)

Now the data is in the proper format, I run the cspade-algorithm with some parameters:

sequences <- cspade(data      = data_for_fseq_mining_trans, 
                    parameter = list(support = 0.3, maxsize = 10, maxlen = 10, mingap = 1, maxgap = 10),
                    control   = list(tidList = TRUE, verbose = TRUE))

Summarizing the results (sequence and relative support):

sequences_df <- cbind(sequence = labels(sequences), support = sequences@quality)

              sequence support
1                  <{A}>     1.0
2                  <{B}>     1.0
3              <{A},{B}>     1.0
4              <{B},{B}>     0.7
5          <{A},{B},{B}>     0.6
6          <{B},{B},{B}>     0.4
7      <{A},{B},{B},{B}>     0.4
8      <{B},{B},{B},{B}>     0.3
9  <{A},{B},{B},{B},{B}>     0.3
10         <{A},{A},{B}>     0.5
11         <{B},{A},{B}>     0.4
12     <{A},{B},{A},{B}>     0.3
13             <{A},{A}>     0.8
14             <{B},{A}>     0.6
15         <{A},{B},{A}>     0.6
16         <{B},{B},{A}>     0.5
17     <{A},{B},{B},{A}>     0.4

That's perfectly fine, but now I would like to know, for each transaction, whether each sequence is present or not (TRUE/FALSE). To do this, I tried to use the tidList:

sequences_score <- as.matrix(sequences@tidLists@data)

      [,1] [,2] [,3]  [,4]  [,5]  [,6]  [,7]  [,8]  [,9] [,10] [,11] [,12] [,13] [,14] [,15] [,16] [,17]
 [1,] TRUE TRUE TRUE  TRUE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE  TRUE  TRUE  TRUE  TRUE
 [2,] TRUE TRUE TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
 [3,] TRUE TRUE TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE FALSE  TRUE  TRUE  TRUE  TRUE  TRUE
 [4,] TRUE TRUE TRUE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE FALSE  TRUE FALSE FALSE FALSE FALSE
 [5,] TRUE TRUE TRUE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE  TRUE  TRUE  TRUE  TRUE FALSE
 [6,] TRUE TRUE TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
 [7,] TRUE TRUE TRUE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE FALSE  TRUE FALSE FALSE FALSE FALSE
 [8,] TRUE TRUE TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE
 [9,] TRUE TRUE TRUE  TRUE  TRUE FALSE FALSE FALSE FALSE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE
[10,] TRUE TRUE TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE

From this result, I assume each row corresponds to a transaction and each column to a sequence. But when looking at the 4th column, it says that pattern '
<{B},{B}>' is not present in transactions 2, 4 and 7. Though these transactions clearly all contain this pattern. Are my assumptions about the output wrong?

An alternative approach is to use this piece of code provided by juliesls: R arulesSequences Find which patterns are supported by a sequence

When applying the following lines of code, an error occurs.

ids             <- unique(data_for_fseq_mining_trans@itemsetInfo$sequenceID)
sequences_score <- data.frame()

for (seq_id in 1:length(sequences)){
  sequences_score[,labels(sequences[seq_id])] <- logical(0)
}

for (id in ids){
  transaction_subset    <- data_for_fseq_mining_trans[data_for_fseq_mining_trans@itemsetInfo$sequenceID==id]
  sequences_score[id, ] <- as.logical(support(x = sequences, transactions = 
  transaction_subset, type="absolute"))
}

Any clues?

Answer 1

To see whether each sequence is present or not you can indeed use your provided code:

sequences_score <- as.matrix(sequences@tidLists@data)

However, you have to map the resulting matrix to your data using another property of your sequence object as follows:

# Get mapping ids, change to numeric values
mapping_ids      <- as.numeric(sequences@tidLists@transactionInfo$sequenceID)

# Then map your matrix sequence_score to correspond to the order of your data
sequences_score  <- sequences_score[order(mapping_ids), ]