Association Rule Mining

Association Rule Mining is a Rule-Based Machine Learning method, to analyze dataset for patterns or co-occurrences to create list of Association Rules and is a method for discovering relations between data items, creating Frequent Itemsets. Apriori Algorithm is the most famous example of this method.

Tldr

It's the discovering of statistically significant associations in data.


Algorithms:


Notes

  • Support and confidence are common metrics used to evaluate association rules. Support measures the frequency of a particular itemset, while confidence measures the likelihood of the consequent in the rule being true, given that the antecedent is true.
  • The Apriori algorithm uses a level-wise search strategy to discover frequent item sets and generate association rules. It prunes the search space by applying the downward closure property to avoid generating unnecessary candidate item sets.
  • Association rule mining is widely used in various domains including market basket analysis in retail, Recommender Systems, and web usage mining to discern patterns in user behavior.