apriori {arules}

Mining Associations with the Apriori Algorithm

Description

Mine frequent itemsets, association rules or association hyperedges using the Apriori algorithm.

Usage

apriori(data, parameter = NULL, appearance = NULL, control = NULL, ...)

Arguments

data

object of class transactions. Any data structure which can be coerced into transactions (e.g., a binary matrix, a data.frame or a tibble) can also be specified and will be internally coerced to transactions.

parameter

object of class APparameter or named list. The default behavior is to mine rules with minimum support of 0.1, minimum confidence of 0.8, maximum of 10 items (maxlen), and a maximal time for subset checking of 5 seconds (maxtime).

appearance

object of class APappearance or named list. With this argument item appearance can be restricted (implements rule templates). By default all items can appear unrestricted.

control

object of class APcontrol or named list. Controls the algorithmic performance of the mining algorithm (item sorting, report progress (verbose), etc.)

...

Additional arguments are for convenience added to the parameter list.

Details

The Apriori algorithm (Agrawal et al, 1993) employs level-wise search for frequent itemsets. The used C implementation of Apriori by Christian Borgelt (2003) includes some improvements (e.g., a prefix tree and item sorting).

Warning about automatic conversion of matrices or data.frames to transactions. It is preferred to create transactions manually before calling apriori() to have control over item coding. This is especially important when you are working with multiple datasets or several subsets of the same dataset. To read about item coding, see itemCoding.

If a data.frame is specified as x, then the data is automatically converted into transactions by discretizing numeric data using discretizeDF() and then coercion to transactions. The discretization may fail if the data is not well behaved.

Apriori only creates rules with one item in the RHS (Consequent). The default value in APparameter for minlen is 1. This meains that rules with only one item (i.e., an empty antecedent/LHS) like

\{\} => \{beer\}

will be created. These rules mean that no matter what other items are involved, the item in the RHS will appear with the probability given by the rule's confidence (which equals the support). If you want to avoid these rules then use the argument parameter = list(minlen = 2).

Notes on run time and memory usage: If the minimum support is chosen too low for the dataset, then the algorithm will try to create an extremely large set of itemsets/rules. This will result in very long run time and eventually the process will run out of memory. To prevent this, the default maximal length of itemsets/rules is restricted to 10 items (via the parameter element maxlen = 10) and the time for checking subsets is limited to 5 seconds (via maxtime = 5). The output will show if you hit these limits in the "checking subsets" line of the output. The time limit is only checked when the subset size increases, so it may run significantly longer than what you specify in maxtime. Setting maxtime = 0 disables the time limit.

Interrupting execution with Control-C/Esc is not recommended. Memory cleanup will be prevented resulting in a memory leak. Also, interrupts are only checked when the subset size increases, so it may take some time till the execution actually stops.

Value

Returns an object of class rules or itemsets.

Author(s)

Michael Hahsler and Bettina Gruen

References

R. Agrawal, T. Imielinski, and A. Swami (1993) Mining association rules between sets of items in large databases. In Proceedings of the ACM SIGMOD International Conference on Management of Data, pages 207–216, Washington D.C. doi:10.1145/170035.170072

Christian Borgelt (2012) Frequent Item Set Mining. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery 2(6):437-456. J. Wiley & Sons, Chichester, United Kingdom 2012. doi:10.1002/widm.1074

Christian Borgelt and Rudolf Kruse (2002) Induction of Association Rules: Apriori Implementation. 15th Conference on Computational Statistics (COMPSTAT 2002, Berlin, Germany) Physica Verlag, Heidelberg, Germany.

Christian Borgelt (2003) Efficient Implementations of Apriori and Eclat. Workshop of Frequent Item Set Mining Implementations (FIMI 2003, Melbourne, FL, USA).

APRIORI Implementation: https://borgelt.net/apriori.html

See Also

Other mining algorithms: APappearance-class, AScontrol-classes, ASparameter-classes, eclat(), fim4r(), ruleInduction(), weclat()

Examples

## Example 1: Create transaction data and mine association rules
a_list <- list(
      c("a","b","c"),
      c("a","b"),
      c("a","b","d"),
      c("c","e"),
      c("a","b","d","e")
      )

## Set transaction names
names(a_list) <- paste("Tr",c(1:5), sep = "")
a_list

## Use the constructor to create transactions
trans1 <- transactions(a_list)
trans1

rules <- apriori(trans1)
inspect(rules)

## Example 2: Mine association rules from an existing transactions dataset
##   using different minimum support and minimum confidence thresholds
data("Adult")

rules <- apriori(Adult,
	parameter = list(supp = 0.5, conf = 0.9, target = "rules"))
summary(rules)

# since ... gets automatically added to parameter, we can also write the 
#  same call shorter:
apriori(Adult, supp = 0.5, conf = 0.9, target = "rules")