AScontrol-classes: Classes AScontrol, APcontrol, ECcontrol - Specifying the...

In mhahsler/arules: Mining Association Rules and Frequent Itemsets

Classes AScontrol, APcontrol, ECcontrol — Specifying the control Argument of Apriori and Eclat

Description

The AScontrol class holds the algorithmic parameters for the used mining algorithms. APcontrol and ECcontrol directly extend AScontrol with additional slots for parameters only suitable for the algorithms Apriori (APcontrol) and Eclat (ECcontrol).

Slots

sort

an integer scalar indicating how to sort items with respect to their frequency: (default: 2)

• 1: ascending

• -1: descending

• 0: do not sort

• 2: ascending

• -2: descending with respect to transaction size sum

verbose

a logical indicating whether to report progress

filter

a numeric scalar indicating how to filter unused items from transactions (default: 0.1)

• =0: do not filter items with respect to. usage in sets

• <0: fraction of removed items for filtering

• >0: take execution times ratio into account

tree

a logical indicating whether to organize transactions as a prefix tree (default: TRUE)

heap

a logical indicating whether to use heapsort instead of quicksort to sort the transactions (default: TRUE)

memopt

a logical indicating whether to minimize memory usage instead of maximize speed (default: FALSE)

load

a logical indicating whether to load transactions into memory (default: TRUE)

sparse

a numeric value for the threshold for sparse representation (default: 7)

Available Slots by Subclass

• APcontrol: filter, tree, heap, memopt, load, sort, verbose

• ECcontrol: sparse, sort, verbose

Objects from the Class

A suitable default control object will be automatically created by the apriori() or the eclat() function. By specifying a named list (names equal to slots) as the control argument for apriori() or eclat(), default values can be replaced with the values in the list.

Objects can also be created via coercion.

Coercions

• as("NULL", "APcontrol")

• as("list", "APcontrol")

• as("NULL", "ECcontrol")

• as("list", "ECcontrol")

Author(s)

Michael Hahsler and Bettina Gruen

References

Christian Borgelt (2004) Apriori — Finding Association Rules/Hyperedges with the Apriori Algorithm. https://borgelt.net/apriori.html

See Also

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

mhahsler/arules documentation built on March 19, 2024, 5:45 p.m.