Association rule learning

Algorithm

Association rule learning is used to discover interesting correlations in large bodies of data. For example, a supermarket would use it to discover which items are normally bought together so it can place them together on the shelves: beer is normally bought at the same time as crisps or nuts (and, according to one report, nappies, presumably because people who have babies are more likely to drink at home :-)). In another domain, an analysis of visible building faults that typically occur together could be used to determine when it makes economic sense to perform expensive tests for related concealed problems.

Apriori, Eclat and FPGrowth are all specific association rule learning algorithms.

alias
subtype
Apriori Eclat FPGrowth
has functional building block
FBB_Feature discovery
has input data type
IDT_Vector of categorical variables
has internal model
has output data type
ODT_Vector of categorical variables ODT_Probability
has learning style
LST_Unsupervised
has parametricity
PRM_Nonparametric with hyperparameter(s)
has relevance
REL_Relevant
uses
sometimes supports
mathematically similar to