Association Rule Mining: Definitions

The association rules can be classified into two categories:
Boolean Association Rules
For example, the information that a customer who purchases a keyboard also tends to buy a mouse is represented in an association rule: 
  Keyboard ⇒ Mouse [support=6%, confidence=70%]
where

  • support(A⇒B) = NAB/N = the percentage of task-relevant data transactions for which the pattern is true,

  • confidence(A⇒B) = NAB/NA = the measure of certainty or trustworthiness associated with each discovered pattern,

  • N = total number of tuples,

  • NA = number of tuples containing keyboard, and

  • NAB = number of tuples containing both keyboard and mouse.

Quantitative Association Rules
For example, the information that a person who is between 26 to 30 years old also owns 1 or 2 cars is represented in association rule: 
  (Age=26..30) ⇒ (Cars=1,2) [Support=3%, confidence=36%]
where

  • support(A⇒B) = NAB/N,

  • confidence(A⇒B) = NAB/NA,

  • N = total number of tuples,

  • NA = number of tuples containing persons aged from 26 to 30, and

  • NAB = number of tuples containing persons each of whom is 26 to 30 years old also owns 1 or 2 cars.



      I walked into the gym and see a bunch of ladies working out,    
      I ask the guy who is running the gym,    
      “Sir, what machine should I use to impress the ladies?”    
      He smiles says “Try the ATM in the lobby.”